release: bump version to 0.60.0

release: bump version to 0.59.0
release: bump version to 0.58.0
2026-05-17 02:50:42 +08:00 · 2026-05-16 05:02:05 +08:00 · 2026-05-15 17:40:07 +08:00 · 2026-05-14 01:02:17 +08:00 · 2026-05-13 18:21:03 +08:00 · 2026-05-13 15:31:29 +08:00
164 changed files with 15444 additions and 1872 deletions
--- a/.dockerignore
+++ b/.dockerignore
@@ -2,9 +2,14 @@

 !pyproject.toml
 !uv.lock
+!VERSION
+!backend/
+!backend/**
 !aiprovider/
 !aiprovider/**

+backend/.env
+backend/.env.*
 aiprovider/.env
 aiprovider/.env.*
 !aiprovider/.env.example
--- a/.gitea/workflows/ci.yaml
+++ b/.gitea/workflows/ci.yaml
@@ -0,0 +1,64 @@
+name: ci
+
+on:
+  push:
+    branches:
+      - dev
+      - main
+  pull_request:
+
+env:
+  REGISTRY: gitea.rclaw.top
+  IMAGE_NAMESPACE: linkong/planet
+
+jobs:
+  backend:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install uv
+        run: curl -LsSf https://astral.sh/uv/install.sh | sh
+      - name: Sync Python dependencies
+        run: ~/.local/bin/uv sync --group dev
+      - name: Run backend smoke tests
+        working-directory: backend
+        run: PYTHONPATH=. "$GITHUB_WORKSPACE/.venv/bin/python" -m pytest -s tests/test_api.py tests/test_realtime_sources.py -q
+
+  frontend:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install Bun
+        run: curl -fsSL https://bun.sh/install | bash
+      - name: Build frontend
+        working-directory: frontend
+        run: |
+          ~/.bun/bin/bun install --frozen-lockfile
+          ~/.bun/bin/bun run build
+
+  delivery:
+    runs-on: ubuntu-latest
+    needs:
+      - backend
+      - frontend
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install Helm
+        run: |
+          mkdir -p "$HOME/.local/bin"
+          curl -fsSL https://get.helm.sh/helm-v3.15.4-linux-amd64.tar.gz -o /tmp/helm.tar.gz
+          tar -xzf /tmp/helm.tar.gz -C /tmp
+          mv /tmp/linux-amd64/helm "$HOME/.local/bin/helm"
+          echo "$HOME/.local/bin" >> "$GITHUB_PATH"
+      - name: Docker build smoke
+        run: |
+          docker build -t "$REGISTRY/$IMAGE_NAMESPACE/frontend:${GITHUB_SHA}" ./frontend
+          docker build -t "$REGISTRY/$IMAGE_NAMESPACE/backend:${GITHUB_SHA}" -f backend/Dockerfile .
+          docker build -t "$REGISTRY/$IMAGE_NAMESPACE/aiprovider:${GITHUB_SHA}" -f aiprovider/Dockerfile .
+      - name: Helm template smoke
+        run: |
+          helm lint deploy/helm/planet
+          helm template planet-staging deploy/helm/planet \
+            --namespace planet-staging \
+            -f deploy/helm/planet/values.single-node.yaml \
+            --set image.tag="${GITHUB_SHA}" >/tmp/planet-rendered.yaml
--- a/.gitea/workflows/deploy-staging.yaml
+++ b/.gitea/workflows/deploy-staging.yaml
@@ -0,0 +1,67 @@
+name: deploy-staging
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+
+env:
+  REGISTRY: gitea.rclaw.top
+  IMAGE_NAMESPACE: linkong/planet
+  RELEASE_NAME: planet-staging
+  NAMESPACE: planet-staging
+
+jobs:
+  deploy:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install deploy tools
+        run: |
+          mkdir -p "$HOME/.local/bin"
+          curl -fsSL https://get.helm.sh/helm-v3.15.4-linux-amd64.tar.gz -o /tmp/helm.tar.gz
+          tar -xzf /tmp/helm.tar.gz -C /tmp
+          mv /tmp/linux-amd64/helm "$HOME/.local/bin/helm"
+          curl -fsSL https://dl.k8s.io/release/v1.30.5/bin/linux/amd64/kubectl -o /tmp/kubectl
+          install -m 0755 /tmp/kubectl "$HOME/.local/bin/kubectl"
+          echo "$HOME/.local/bin" >> "$GITHUB_PATH"
+      - name: Configure kubeconfig
+        run: |
+          mkdir -p "$HOME/.kube"
+          printf "%s" "${{ secrets.KUBE_CONFIG_STAGING }}" | base64 -d > "$HOME/.kube/config"
+          chmod 600 "$HOME/.kube/config"
+      - name: Deploy Helm release
+        run: |
+          kubectl create namespace "$NAMESPACE" --dry-run=client -o yaml | kubectl apply -f -
+          helm upgrade --install "$RELEASE_NAME" deploy/helm/planet \
+            --namespace "$NAMESPACE" \
+            -f deploy/helm/planet/values.single-node.yaml \
+            --set global.imageRegistry="$REGISTRY" \
+            --set global.imageNamespace="$IMAGE_NAMESPACE" \
+            --set image.tag="${GITHUB_SHA}"
+      - name: Wait for rollout
+        run: |
+          kubectl rollout status deployment/planet-frontend -n "$NAMESPACE" --timeout=180s
+          kubectl rollout status deployment/planet-backend -n "$NAMESPACE" --timeout=180s
+          kubectl rollout status deployment/planet-aiprovider -n "$NAMESPACE" --timeout=180s
+      - name: Smoke test services
+        run: |
+          kubectl run planet-smoke-${GITHUB_RUN_NUMBER} \
+            --rm -i --restart=Never \
+            --namespace "$NAMESPACE" \
+            --image=curlimages/curl:8.11.1 \
+            --command -- sh -c '
+              set -eu
+              curl -fsS http://planet-frontend:3000/ >/dev/null
+              curl -fsS http://planet-frontend:3000/health >/dev/null
+              curl -fsS http://planet-frontend:3000/api/health >/dev/null
+              curl -fsS http://planet-backend:8000/health >/dev/null
+              curl -fsS http://planet-aiprovider:8010/health >/dev/null
+            '
+      - name: Collect diagnostics on failure
+        if: failure()
+        run: |
+          kubectl get all -n "$NAMESPACE" -o wide || true
+          kubectl describe pods -n "$NAMESPACE" || true
+          kubectl logs -n "$NAMESPACE" -l app.kubernetes.io/instance="$RELEASE_NAME" --all-containers --tail=200 || true
--- a/.gitea/workflows/release.yaml
+++ b/.gitea/workflows/release.yaml
@@ -0,0 +1,58 @@
+name: release
+
+on:
+  push:
+    branches:
+      - main
+    tags:
+      - "v*"
+  workflow_dispatch:
+
+env:
+  REGISTRY: gitea.rclaw.top
+  IMAGE_NAMESPACE: linkong/planet
+
+jobs:
+  images:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Resolve image tags
+        id: meta
+        run: |
+          echo "sha_tag=${GITHUB_SHA}" >> "$GITHUB_OUTPUT"
+          if printf "%s" "${GITHUB_REF}" | grep -q '^refs/tags/v'; then
+            echo "release_tag=${GITHUB_REF_NAME}" >> "$GITHUB_OUTPUT"
+          else
+            echo "release_tag=" >> "$GITHUB_OUTPUT"
+          fi
+      - name: Login to registry
+        run: echo "${{ secrets.REGISTRY_PASSWORD }}" | docker login "$REGISTRY" -u "${{ secrets.REGISTRY_USER }}" --password-stdin
+      - name: Build and push images
+        run: |
+          for service in frontend backend aiprovider; do
+            case "$service" in
+              frontend)
+                dockerfile="./frontend/Dockerfile"
+                context="./frontend"
+                ;;
+              backend)
+                dockerfile="backend/Dockerfile"
+                context="."
+                ;;
+              aiprovider)
+                dockerfile="aiprovider/Dockerfile"
+                context="."
+                ;;
+            esac
+
+            image="$REGISTRY/$IMAGE_NAMESPACE/$service:${{ steps.meta.outputs.sha_tag }}"
+            docker build -t "$image" -f "$dockerfile" "$context"
+            docker push "$image"
+
+            if [ -n "${{ steps.meta.outputs.release_tag }}" ]; then
+              release_image="$REGISTRY/$IMAGE_NAMESPACE/$service:${{ steps.meta.outputs.release_tag }}"
+              docker tag "$image" "$release_image"
+              docker push "$release_image"
+            fi
+          done
--- a/.gitignore
+++ b/.gitignore
@@ -8,6 +8,7 @@
 .env
 .env.local
 .env.*.local
+config/earth-boundary-sources.local.json
 *.pem
 *.key
 *.crt
@@ -150,3 +151,9 @@ temp/
 # Runtime Data
 # ----------------------
 data/ai/bgp-briefs/
+data/earth-boundary-sources/
+
+# Generated Earth boundary tile artifacts. Keep source configs and builders in
+# Git; publish PMTiles/MVT artifacts through release/deploy storage instead of
+# committing thousands of generated tile files.
+frontend/public/earth/data/boundaries/
--- a/README.md
+++ b/README.md
@@ -8,68 +8,54 @@

 ## 系统架构

+当前仓库的核心形态是“Web Earth 可视化 + React 运维台 + FastAPI 数据与 AI 编排后端 + 独立模型适配层”。物理大屏与 UE 客户端仍是长期方向，但不再作为本地开发和当前发布的必需运行单元。
+
 ```
-┌─────────────────────────────────────────────────────────────────────────┐
-│                           物理大屏展示层                                  │
-│  ┌─────────────────────────────────────────────────────────────────┐    │
-│  │         偏振片3D大屏 (2m×3m, 4K, 120Hz, 眼镜式)                 │    │
-│  │  ┌─────────────────────────────────────────────────────────┐    │    │
-│  │  │              虚幻引擎 UE5 客户端                          │    │    │
-│  │  │  ├── 3D地球渲染 (Cesium for UE)                          │    │    │
-│  │  │  ├── 算力点可视化 (GPU集群、智算中心)                     │    │    │
-│  │  │  ├── 连接弧线 (光缆、路由、数据流向)                      │    │    │
-│  │  │  ├── 粒子效果 (数据流动、告警提示)                        │    │    │
-│  │  │  └── 自动巡航相机 + 交互控制                              │    │    │
-│  │  └─────────────────────────────────────────────────────────┘    │    │
-│  └─────────────────────────────────────────────────────────────────┘    │
-└─────────────────────────────────────────────────────────────────────────┘
-                                    ▲
-                                    │ WebSocket (实时推送)
-                                    │ 120Hz 心跳 / 数据帧同步
-                                    ▼
-┌─────────────────────────────────────────────────────────────────────────┐
-│                           数据中台服务层 (FastAPI)                        │
-│  ┌─────────────────────────────────────────────────────────────────┐    │
-│  │                      API Gateway (Redis 限流)                    │    │
-│  └─────────────────────────────────────────────────────────────────┘    │
-│                                    │                                    │
-│  ┌───────────────────┬──────────────────────────┬──────────────────┐   │
-│  │    数据采集服务    │      核心业务服务         │    运维管理服务   │   │
-│  │  ┌─────────────┐  │  ┌─────────────────┐    │  ┌─────────────┐ │   │
-│  │  │ 调度中心    │  │  │ WebSocket 服务  │    │  │ 用户管理    │ │   │
-│  │  │ (Celery)   │  │  │ (FastAPI)       │    │  │ (JWT Auth)  │ │   │
-│  │  └─────────────┘  │  └─────────────────┘    │  └─────────────┘ │   │
-│  │  ┌─────────────┐  │  ┌─────────────────┐    │  ┌─────────────┐ │   │
-│  │  │ 采集器池    │  │  │ 数据查询 API    │    │  │ 数据源配置  │ │   │
-│  │  │ (10+源)    │  │  │ (REST)          │    │  │ 监控告警    │ │   │
-│  │  └─────────────┘  │  └─────────────────┘    │  └─────────────┘ │   │
-│  │  ┌─────────────┐  │  ┌─────────────────┐    │  ┌─────────────┐ │   │
-│  │  │ 消息队列    │  │  │ 态势分析引擎    │    │  │ 系统配置    │ │   │
-│  │  │ (Kafka)    │  │  │ (计算/聚合)     │    │  │ 日志审计    │ │   │
-│  │  └─────────────┘  │  └─────────────────┘    │  └─────────────┘ │   │
-│  └───────────────────┴──────────────────────────┴──────────────────┘   │
-└─────────────────────────────────────────────────────────────────────────┘
-                                    ▲
-                                    │ 内部 API 调用
-                                    ▼
-┌─────────────────────────────────────────────────────────────────────────┐
-│                           Web管理端 (React Admin)                        │
-│  ┌─────────────────────────────────────────────────────────────────┐    │
-│  │  登录页 │ 仪表盘 │ 用户管理 │ 数据源配置 │ 任务监控 │ 系统配置   │    │
-│  └─────────────────────────────────────────────────────────────────┘    │
-└─────────────────────────────────────────────────────────────────────────┘
-                                    ▲
-                                    │ PostgreSQL / Redis
-                                    ▼
-┌─────────────────────────────────────────────────────────────────────────┐
-│                           数据存储层                                      │
-│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐    │
-│  │ PostgreSQL  │  │ TimescaleDB │  │    Redis    │  │  MinIO      │    │
-│  │ (用户/配置) │  │ (时序数据)  │  │ (缓存/会话) │  │ (文件存储) │    │
-│  └─────────────┘  └─────────────┘  └─────────────┘  └─────────────┘    │
-└─────────────────────────────────────────────────────────────────────────┘
+┌─────────────────────────────────────────────────────────────────────┐
+│                         浏览器展示与运维层                            │
+│  ┌──────────────────────────────┐  ┌──────────────────────────────┐ │
+│  │ Web Earth                    │  │ React 运维台                  │ │
+│  │ frontend/public/earth        │  │ frontend/src                  │ │
+│  │ Three.js 地球 / HUD / 新闻    │  │ 数据源 / 告警 / AI 设置       │ │
+│  │ 国界精度 / 品牌内容配置       │  │ 提示词配置 / 用户与系统配置   │ │
+│  └──────────────────────────────┘  └──────────────────────────────┘ │
+└─────────────────────────────────────────────────────────────────────┘
+                       │ REST / WebSocket
+                       ▼
+┌─────────────────────────────────────────────────────────────────────┐
+│                       FastAPI 业务与编排后端                          │
+│  ┌────────────────────┐ ┌────────────────────┐ ┌─────────────────┐ │
+│  │ 数据 API 与认证     │ │ Earth 新闻增强      │ │ 告警与态势简报   │ │
+│  │ JWT / 权限 / 审计   │ │ 位置推断 / 本地化   │ │ BGP / 告警研判   │ │
+│  └────────────────────┘ └────────────────────┘ └─────────────────┘ │
+│  ┌────────────────────┐ ┌────────────────────┐ ┌─────────────────┐ │
+│  │ 系统运行配置        │ │ 默认提示词注册表     │ │ 未来 Agent Runtime│ │
+│  │ system_settings     │ │ 代码发布 + DB 覆盖  │ │ 工具/证据/工作流   │ │
+│  └────────────────────┘ └────────────────────┘ └─────────────────┘ │
+└─────────────────────────────────────────────────────────────────────┘
+        │ SQLAlchemy / Redis Stream            │ 纯净 LLM 调用
+        ▼                                      ▼
+┌──────────────────────────────┐      ┌──────────────────────────────┐
+│ PostgreSQL / Redis            │      │ aiprovider                   │
+│ 用户、配置、采集结果、新闻     │      │ provider + protocol adapter  │
+│ Stream、缓存、运行状态         │      │ OpenAI / MiniMax / Ollama 等 │
+└──────────────────────────────┘      └──────────────────────────────┘
+        ▲
+        │ 采集器 / 外部数据源
+        ▼
+┌─────────────────────────────────────────────────────────────────────┐
+│ RSS 新闻、BGP 观测、公开数据源、后续 WebSearch/OCR/语音识别等工具      │
+└─────────────────────────────────────────────────────────────────────┘
 ```

+架构边界:
+
+- `backend` 负责业务语义、证据收集、提示词选择、AI 任务编排、权限和数据落库。
+- `aiprovider` 只负责把纯净模型请求适配到不同供应商或协议，不内置具体业务提示词。
+- 默认提示词随代码发布并保存在 `backend/app/ai_tasks/default_prompts.json`，运维台可在数据库中保存覆盖值，重置时回到当前代码版本的默认提示词。
+- Earth 新闻保留英文原文，中文展示结果存入 `localizations`，前端默认展示 `zh-CN` 的 `display_title`、`display_summary` 和中文地域/状态文案。
+- Earth LLM 指令、语音识别、多角色态势研判属于后续 Agent Runtime 方向，计划见 [docs/plans/agents-earth-command-runtime-plan.md](/home/ray/dev/linkong/planet/docs/plans/agents-earth-command-runtime-plan.md)。
+
 ## 四大核心要素

 | 层级 | 要素 | 描述 |
@@ -87,11 +73,10 @@
 |------|------|------|
 | FastAPI | 0.109+ | Web 框架 |
 | SQLAlchemy | 2.0+ | ORM |
-| Alembic | - | 数据库迁移 |
-| Celery | 5.3+ | 任务队列 |
-| Redis | 7.0+ | 缓存/消息 |
-| Kafka | 3.0+ | 事件流 |
+| uv | - | Python 依赖与命令运行 |
+| Redis | 7.0+ | 缓存、Stream 与运行协调 |
 | PyJWT | - | 认证 |
+| APScheduler / 后台任务 | - | 采集、增强与运行时任务 |

 ### 前端 (React Admin)

@@ -102,6 +87,7 @@
 | Axios | HTTP 客户端 |
 | Socket.io-client | WebSocket 客户端 |
 | ECharts | 统计图表 |
+| Three.js | Earth 3D 地球渲染 |
 | Bun | 前端包管理与脚本运行 |

 前端工程统一使用 Bun：
@@ -110,22 +96,16 @@
 - 运行脚本使用 `bun run <script>`
 - 不使用 `npm`、`pnpm`、`yarn`

-### 虚幻引擎客户端
+### 大屏与 3D 展示方向

-| 组件 | 版本 | 用途 |
-|------|------|------|
-| Unreal Engine 5 | 5.3+ | 3D 渲染引擎 |
-| Cesium for Unreal | 1.5+ | 地理可视化 |
-| Niagara | - | 粒子系统 |
+当前发布优先使用浏览器 Web Earth。UE5 / Cesium for Unreal / Niagara 可作为后续物理大屏方向接入，但不是本地开发闭环的必需组件。

 ### 数据库

 | 组件 | 用途 |
 |------|------|
 | PostgreSQL 15+ | 关系数据 |
-| TimescaleDB | 时序数据扩展 |
-| Redis 7+ | 缓存/会话 |
-| MinIO | S3 兼容存储 |
+| Redis 7+ | 缓存、Stream、运行状态 |

 ### 部署

@@ -152,9 +132,9 @@
 | P0 | Epoch AI | 每小时 |
 | P0 | Hugging Face | 每 2 小时 |
 | P0 | GitHub | 每 4 小时 |
-| P0 每日 |
+| P0 | 海底光缆 / IXP / 卫星等基础设施数据 | 每日或按源刷新 |
 | P0 | PeeringDB | 每 2 小时 |
-| P1 | Cloudflare Radar | | TeleGeography | 每小时 |
+| P1 | Cloudflare Radar / TeleGeography | 每小时 |
 | P1 | CAIDA BGPStream | 每 15 分钟 |

 ## 项目结构
@@ -166,20 +146,18 @@
 │   │   ├── core/         # 核心配置
 │   │   ├── models/       # 数据模型
 │   │   ├── schemas/      # Pydantic 模型
-│   │   ├── services/     # 业务逻辑
-│   │   └── tasks/        # Celery 任务
+│   │   ├── services/     # 业务逻辑与 AI 任务编排
+│   │   └── ai_tasks/     # 默认提示词与 AI 任务定义
 │   └── tests/
+├── aiprovider/            # 独立模型供应商适配层
 ├── frontend/              # React 管理后台
 │   ├── src/
 │   │   ├── components/   # 组件
 │   │   ├── pages/        # 页面
 │   │   ├── services/     # API 服务
 │   │   └── store/        # 状态管理
-│   └── tests/
-├── unreal/                # UE5 大屏客户端
-│   ├── Content/
-│   ├── Source/
-│   └── Plugins/
+│   ├── public/earth/      # Web Earth 静态应用
+│   └── tests/             # 前端测试
 ├── data/                  # 数据文件
 ├── docs/                  # 文档
 ├── scripts/               # 脚本
@@ -244,7 +222,7 @@ bun run build
 ./planet.sh start --allow-lan
 ```

-这会让前端监听 `0.0.0.0:3000`，后端监听 `0.0.0.0:8000`。
+这会让前端监听 `0.0.0.0:3000`，后端监听 `0.0.0.0:8000`，AI Provider 通过 Docker 发布到 `0.0.0.0:8010`。启动前脚本会检查这三个端口；如果 WSL/Linux 侧无法释放端口，并检测到 Windows 侧 listener 或旧 `portproxy`，会请求管理员 PowerShell 清理。

 ### 2. 先确认 WSL 内部服务正常

@@ -253,14 +231,16 @@ bun run build
 ```bash
 curl http://localhost:3000
 curl http://localhost:8000/health
-ss -ltnp | grep -E ':3000|:8000'
+curl http://localhost:8010/health
+ss -ltnp | grep -E ':3000|:8000|:8010'
 ```

 预期：

 - `3000` 返回前端 HTML
 - `8000/health` 返回健康检查 JSON
- `ss` 中能看到 `0.0.0.0:3000` 和 `0.0.0.0:8000`
+- `8010/health` 返回 AI Provider 健康检查 JSON
+- `ss` 中能看到 `0.0.0.0:3000`、`0.0.0.0:8000` 和 `0.0.0.0:8010`，或 Docker 已发布 `8010`

 如果这一步不通，先不要继续做 Windows 转发。

@@ -271,42 +251,31 @@ ss -ltnp | grep -E ':3000|:8000'
 ```powershell
 curl http://localhost:3000
 curl http://localhost:8000/health
+curl http://localhost:8010/health
 ```

 在常见的 WSL2 开发环境下，Windows 通常可以直接通过 `localhost` 访问 WSL 中的服务。

-### 4. 如果需要让局域网设备访问，再做 Windows 端口转发
+### 4. 如果需要让局域网设备访问，清理端口和防火墙

-注意：下面的命令必须在“以管理员身份运行”的 PowerShell 中执行。
+`./planet.sh start --allow-lan` 不再启动额外的 Windows 端口转发进程。它直接让开发服务对 `3000` / `8000` / `8010` 开放，并在启动前尝试释放这些端口。端口被 Windows 侧 listener 或旧 `portproxy` 占用时，脚本会请求一次管理员 PowerShell 清理。

-先把 Windows 对外网卡上的 `3000` / `8000` 转发到 Windows 本机 `127.0.0.1`：
+如果以前手动配置过持久 `portproxy`，若自动请求被取消，可以手动清理，避免 `iphlpsvc` 继续占用端口：

 ```powershell
 netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=3000
 netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8000
-
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=3000 connectaddress=127.0.0.1 connectport=3000
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=8000 connectaddress=127.0.0.1 connectport=8000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8010
 ```

-再放行 Windows 防火墙：
+脚本会检测 Windows 防火墙是否已放行 `3000` / `8000` / `8010`。如果缺少规则，会触发一次 Windows UAC 管理员 PowerShell 请求来自动创建。若自动请求被取消，也可以手动执行：

 ```powershell
 New-NetFirewallRule -DisplayName "WSL Planet 3000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3000
 New-NetFirewallRule -DisplayName "WSL Planet 8000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8000
+New-NetFirewallRule -DisplayName "WSL Planet 8010" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8010
 ```

-检查转发规则是否生效：
-
-```powershell
-netsh interface portproxy show all
-```
-
-预期能看到：
-
- `0.0.0.0:3000 -> 127.0.0.1:3000`
- `0.0.0.0:8000 -> 127.0.0.1:8000`
-
 ### 5. 查 Windows 局域网 IP，并让其他设备访问

 在 Windows PowerShell 中执行：
@@ -321,6 +290,8 @@ ipconfig

 - `http://<Windows局域网IP>:3000/earth`
 - `http://<Windows局域网IP>:3000/admin`
+- `http://<Windows局域网IP>:8000/health`
+- `http://<Windows局域网IP>:8010/health`

 例如：

@@ -329,7 +300,7 @@ ipconfig
 ### 6. 常见现象与判断

 - WSL 中 `curl localhost:3000` 能通，但 Windows 访问 `WSL 的局域网 IP:3000` 不通：这是正常现象之一，优先验证 Windows 的 `localhost:3000`
- Windows `localhost:3000` 能通，但局域网设备访问 `Windows 局域网 IP:3000` 不通：通常缺少 `portproxy` 或防火墙放行
+- Windows `localhost:3000` 能通，但局域网设备访问 `Windows 局域网 IP:3000` 不通：通常是 Windows 防火墙、网络配置或旧 `portproxy` 残留
 - `whoami /groups` 中 `S-1-5-32-544` 显示 `deny only`：说明当前 PowerShell 不是提权管理员窗口

 ### 7. 本项目一次性验证顺序
@@ -338,10 +309,12 @@ ipconfig

 1. WSL 中执行 `curl http://localhost:3000`
 2. WSL 中执行 `curl http://localhost:8000/health`
-3. Windows 中执行 `curl http://localhost:3000`
-4. Windows 中执行 `curl http://localhost:8000/health`
-5. 管理员 PowerShell 配置 `portproxy` 和防火墙
-6. 用手机或其他电脑访问 `http://<Windows局域网IP>:3000/earth`
+3. WSL 中执行 `curl http://localhost:8010/health`
+4. Windows 中执行 `curl http://localhost:3000`
+5. Windows 中执行 `curl http://localhost:8000/health`
+6. Windows 中执行 `curl http://localhost:8010/health`
+7. 按脚本提示完成 Windows 防火墙或端口清理 UAC 请求
+8. 用手机或其他电脑访问 Windows 对外端口，例如 `http://<Windows局域网IP>:3000/earth`

 ## 启动容错参数

@@ -365,18 +338,27 @@ DATABASE_RETRY_INTERVAL=10 \
 - `AI_PROVIDER_START_MAX_RETRIES` / `AI_PROVIDER_RETRY_INTERVAL`: 控制 `aiprovider` 的构建/启动与容器重启自愈，默认 `3` 次、`5` 秒
 - `BACKEND_MAX_RETRIES`: 控制后端进程启动重试次数，默认 `3`
 - `FRONTEND_MAX_RETRIES`: 控制前端 dev server 启动重试次数，默认 `3`
- `BACKEND_HEALTH_CHECK_ATTEMPTS` / `BACKEND_HEALTH_CHECK_INTERVAL`: 控制后端 HTTP 健康检查等待次数与间隔，默认 `10` 次、`2` 秒
+- `BACKEND_HEALTH_CHECK_ATTEMPTS` / `BACKEND_HEALTH_CHECK_INTERVAL`: 控制后端 HTTP 健康检查等待次数与间隔，默认 `60` 次、`2` 秒
 - `FRONTEND_HEALTH_CHECK_ATTEMPTS` / `FRONTEND_HEALTH_CHECK_INTERVAL`: 控制前端 HTTP 可访问检查等待次数与间隔，默认 `10` 次、`2` 秒
 - `AI_PROVIDER_HEALTH_CHECK_ATTEMPTS` / `AI_PROVIDER_HEALTH_CHECK_INTERVAL`: 控制 `aiprovider` HTTP 健康检查等待次数与间隔，默认 `10` 次、`2` 秒

-## AI 接口预留
+## AI 与智能体接口

-项目现在采用“两层”设计:
+项目现在采用“三段式”边界:

- 主后端暴露稳定业务接口: `GET /api/v1/ai/provider/status`、`POST /api/v1/ai/situational-awareness/analyze`
- 独立 `aiprovider` 服务负责适配具体模型供应商
+- `backend`: 暴露业务接口，负责选择任务提示词、组织证据、调用工具、保存 AI 设置和结果。
+- `aiprovider`: 暴露模型网关接口，只负责 provider / protocol 适配，不写入 BGP、新闻、告警等业务提示词。
+- 模型供应商: OpenAI 兼容、MiniMax、Anthropic、Ollama 或其他兼容网关。

-这样前端和业务代码不直接依赖 OpenAI、本地模型网关或其他订阅服务，后续切换部署方式只需要调整环境变量。
+这样前端和业务代码不直接依赖某个模型供应商，后续增加 Agent Runtime、Earth 一键 LLM 指令、语音识别或多角色态势研判时，也可以把业务工作流放在后端，而不是污染模型适配层。
+
+当前已落地的 AI 配置能力:
+
+- 运维台 AI 设置可维护 provider、模型、协议、超时、token 等运行配置。
+- 运维台 AI 设置中的“提示词”页可选择不同功能入口，手动覆盖提示词，并一键重置到默认值。
+- 默认提示词随代码发布，位于 [backend/app/ai_tasks/default_prompts.json](/home/ray/dev/linkong/planet/backend/app/ai_tasks/default_prompts.json)。
+- 覆盖值保存在数据库运行配置中，升级代码后可继续保留现场配置，也可重置到新版本默认提示词。
+- 态势摘要、告警研判、新闻本地化等入口应使用各自任务提示词；调用 `aiprovider` 时只传递当前任务所需的 `prompt` / `system_prompt`。

 主后端建议配置:

@@ -389,39 +371,33 @@ AI_PROVIDER_TIMEOUT_SECONDS=60
 `aiprovider` 服务建议配置:

 ```env
-AI_PROVIDER=openai_compatible
-AI_BASE_URL=https://api.openai.com/v1
+AI_PROVIDER=minimax
+AI_PROVIDER_API=anthropic-messages
+AI_BASE_URL=https://api.minimaxi.com/anthropic
 AI_API_KEY=your_api_key
-AI_MODEL=gpt-4o-mini
+AI_MODEL=MiniMax-M2.7
 AI_TIMEOUT_SECONDS=60
 AI_PROVIDER_SERVICE_TOKEN=change_me
 ```

-OpenAI 兼容场景推荐使用:
+推荐映射关系:

- `AI_PROVIDER=openai_compatible`
+- `vLLM` / `LM Studio` / `One API`: `AI_PROVIDER=openai` + `AI_PROVIDER_API=openai-completions`
+- `MiniMax`: `AI_PROVIDER=minimax` + `AI_PROVIDER_API=anthropic-messages`
+- Claude 兼容网关: `AI_PROVIDER=anthropic` + `AI_PROVIDER_API=anthropic-messages`
+- `Ollama`: `AI_PROVIDER=ollama` + `AI_PROVIDER_API=ollama-generate`

-Claude 兼容场景推荐使用:
-
- `AI_PROVIDER=anthropic`
- `AI_PROVIDER=anthropic_compatible`
- `AI_PROVIDER=claude_compatible`
-
-Ollama 原生场景推荐使用:
-
- `AI_PROVIDER=ollama`
-
-比如 MiniMax 或其他 Claude 兼容网关，可以这样配置:
+比如 MiniMax 可以这样配置:

 ```env
-AI_PROVIDER=claude_compatible
-AI_BASE_URL=https://your-claude-compatible-endpoint.example.com
+AI_PROVIDER=minimax
+AI_PROVIDER_API=anthropic-messages
+AI_BASE_URL=https://api.minimaxi.com/anthropic
 AI_API_KEY=your_api_key
-AI_MODEL=your-claude-compatible-model
+AI_MODEL=MiniMax-M2.7
 AI_TIMEOUT_SECONDS=60
 AI_MAX_TOKENS=1200
 AI_ANTHROPIC_VERSION=2023-06-01
-AI_PROVIDER_SERVICE_TOKEN=change_me
 ```

 如果你要本地直接起模型适配层，项目里已经补了模板:
@@ -429,12 +405,6 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 - [aiprovider/.env.example](/home/ray/dev/linkong/planet/aiprovider/.env.example)
 - [docker-compose.local-model.yml](/home/ray/dev/linkong/planet/docker-compose.local-model.yml)

-推荐映射关系:
-
- `vLLM` / `LM Studio` / `One API`: `AI_PROVIDER=openai_compatible`
- `MiniMax` / Claude 兼容网关: `AI_PROVIDER=claude_compatible`
- `Ollama`: `AI_PROVIDER=ollama`
-
 运行与调用补充:

 - `./planet.sh start` 默认会启动 `aiprovider`
@@ -444,11 +414,13 @@ AI_PROVIDER_SERVICE_TOKEN=change_me

 详细文档:

- [docs/technical/agents-aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/agents-aiprovider.md)
+- [docs/technical/zh/agents-aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/zh/agents-aiprovider.md)
+- [docs/technical/en/agents-aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/en/agents-aiprovider.md)
 - [aiprovider/README.md](/home/ray/dev/linkong/planet/aiprovider/README.md)
 - [docs/technical/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
 - [docs/plans/frontend-ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/plans/frontend-ai-playground-development-plan.md)
 - [docs/plans/agents-situational-awareness-foundation-plan.md](/home/ray/dev/linkong/planet/docs/plans/agents-situational-awareness-foundation-plan.md)
+- [docs/plans/agents-earth-command-runtime-plan.md](/home/ray/dev/linkong/planet/docs/plans/agents-earth-command-runtime-plan.md)

 ## 前端页面布局规范

--- a/TODO.md
+++ b/TODO.md
@@ -1,45 +1,83 @@
 # TODO

- [x] 把 BGP 观测站和异常点的 `hover/click` 手感再磨细一点
- [x] 开始做 BGP 异常和海缆/区域的关联展示
- [x] 做 Earth 侧的 `BGP activity layer`，让低 incident 密度时地图仍然有持续可感知的观测存在感
- [x] 给 Earth BGP 补三层状态表达：`平稳观测态 / 局部波动态 / 事件活跃态`
- [x] 把“当前无活跃事件”改造成“观测网络仍在运行、当前未发现聚合级事件”的状态表达
- [x] 做 collector / region 近 15 分钟 activity score 聚合接口或动态聚合逻辑
- [x] 把 Earth 的 BGP incident 改成 `紧凑事件核 + 向外扩张环形 pulse`，替换当前大面积 glow
- [x] 为 BGP incident 建立符号系统：按事件类型用不同 marker，而不是都用同一种亮点
- [x] 把 incident 地理定位从 `collector-centric` 改成 `prefix-centric`，优先使用 `prefix_geography`，其次 `prefix_scope`，再次 ASN 区域，最后才回退到观测区域质心
- [x] 新增 `prefix_geography` 数据层，不再把 `prefix_scope` 当成 prefix 地理归属本身
- [x] 接入 `IPtoASN / IPtoCountry` 作为 prefix-centric geography 的主数据源
- [x] 接入 `OpenGeoFeed` 作为 prefix geography 的高质量覆盖/override 数据源
- [x] 把 RIR delegated 设计成 prefix geography 的 fallback，而不是主来源
- [ ] 为 `aiprovider` 建立 `provider -> api adapter -> compat policy` 的配置中心，优先落成 `json` 或 `yaml` 文件，运行时按 `provider/model` 读取兼容设置，而不是把专项兼容继续散落在 Python 分支里
- [ ] 为市面上主流 AI 服务补专项兼容配置并固化到配置文件中，至少覆盖 `OpenAI / Anthropic / MiniMax / Ollama / Moonshot / DeepSeek / Qwen / GLM / Gemini / OpenRouter / vLLM / LM Studio / One API`
- [ ] 在兼容配置中补齐可声明项：`api adapter`、`base_url pattern`、`auth header`、`thinking default`、`reasoning block mapping`、`stream path`、`tool-call capability`、`multimodal capability`、`provider-specific request patch`
- [ ] 接入 `inetnum` / `inet6num` whois 作为比 RIR 更细粒度的后备层
- [x] 在 activity layer 之后继续补 `route leak` 和 `path instability / flap` detector
- [ ] 对 [frontend/public/earth/js/bgp.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp.js) 做按职责拆分的小重构，拆成 data / markers / overlays / animation，降低后续维护复杂度
- [ ] 可选优化（非必做）：将 BGP incident/collector 标点改为 HTML marker（参考 worldmonitor 的 `htmlElementsData` 思路），实现近乎固定屏幕尺寸与更高密度可点击性
- [ ] 保持 Earth 当前这批纯个人偏好设置继续走本地持久化：`旋转模式`、HUD 面板显示/隐藏、`地形透明度` 暂不升级到后端系统设置，避免把设备级偏好过早做成全局配置
- [ ] 如果后续明确需要“账号级同步 Earth 偏好”，再单独设计 `Earth user preferences`：优先按用户维度而不是全局系统设置保存，并规划 `localStorage -> backend` 的平滑迁移策略
- [ ] 为 Planet / Earth 补一个可用的日志查看系统：先明确前后端/AI Provider/采集任务的日志入口、最近日志聚合、筛选与 tail 能力，再决定是先做脚本级统一入口还是控制台内置日志面板
- [ ] 重写控制台 UI，逐步抛弃 Ant Design，建立自有组件体系，并统一采用 `tabler.io` / Tabler Icons 作为控制台主图标库
- [ ] 把 Earth 态势新闻源从 [earth_news.py](/home/ray/dev/linkong/planet/backend/app/services/earth_news.py) 的硬编码列表抽成可配置目录，优先保持当前“实时聚合”链路不变，只先解决新闻源不可配置的问题
- [ ] 为 Earth 态势新闻设计后续采集器化方案：明确新闻数据模型、去重策略、区域映射、过期清理和 Earth/AI 复用方式，再决定何时把新闻从实时抓取升级成正式 collector
- [ ] AIS v3.1：修复船只聚合完整性，`/geo/vessels` 合并 raw observation 聚合结果与 legacy `vessel_position + vessel_static` 最新结果，确保 BarentsWatch-only 船只不会因为 AISStream 子集存在而消失，并增加 raw/legacy/final unique MMSI 诊断统计
- [ ] AIS v3.2：把 AISStream 从收满 `max_messages` 后结束的批采集改成长连接 streaming service，持续写入 raw observations，通过内部 `/ws` 的 `vessels` channel 推送新船、位置和航向增量，Earth 前端按 MMSI upsert marker
- [ ] AIS v3.3：修正 AISStream 采集页面状态语义，使用 connecting/streaming/reconnecting/stopped 与 indeterminate 状态，展示运行时长、消息数、unique MMSI、message rate、最近消息和错误，不再用一次性 REST 进度条表示长连接
- [ ] AIS v3.4：修复船只身份字段和名称聚合，MMSI/IMO/callsign 按字符串显示且不带千分位符；查询并列出所有仍以 MMSI 号码或 `MMSI <number>` 作为船名的记录，标注来源、最近观测、message types 和缺失原因，并把这批 fallback-name 船只纳入名称聚合修复集合
- [ ] Earth Live Sync：建立统一态势实时同步链路，新增 `earth_summary` WS channel，任意采集器成功后广播轻量 summary invalidation，前端收到后重新拉 `/api/v1/visualization/geo/summary` 并更新 HUD；同时为 BGP 增加 `bgp` WS channel，使 BGP incidents/anomalies/collectors 在不刷新页面时也能 upsert 图层；卫星采集完成后触发 summary 刷新，必要时按 TLE 版本重新 hydrate 卫星数据
- [ ] AIS v4：开放船只多源聚合策略配置，支持 source priority、字段级规则、freshness 窗口和高级保护开关；保存时校验未知字段、非法模式和危险动态字段锁定，并在聚合接口返回命中的配置版本
- [ ] AIS v5：实现船舶资料 enrichment 与冲突治理，按 `mmsi + imo + name + callsign` 异步补充船型细分、AIS 大类、旗国、尺寸、建造年份、运营方和图片缓存；详情面板展示缓存资料和字段来源，不在实时 AIS 请求链路现场抓第三方页面
- [ ] 为 Earth 地球表面增加一层与基础纹理对齐的材质/纹理 overlay，并在同层叠加国界轮廓参考线；要求国界线与底图稳定对齐，且 hover 到国家轮廓时能高亮当前国家，便于校准地表和增强交互
- [ ] 把 Earth 新闻接入通用巡航队列：按新闻发生地和时间排序生成巡航目标，巡航聚焦到新闻事件时显示对应新闻卡片，并保持实现边界为“通用巡航层 + 新闻业务适配层”，不要再把新闻逻辑直接耦合回 `main.js` 状态机
- [ ] 为未知位置的算力中心建立分层坐标补全链路：优先 `精确坐标 > 站点/园区命中 > 城市 > 州/省 > 国家内主要算力城市 > 国家质心`，并把每次回退的 `confidence / reason / precision` 明确写进统一 GeoJSON
- [ ] 为算力中心补一份可维护的本地位置注册表，例如 `canonical_name / aliases / operator / country / region / city / lat / lon / confidence / source_note`，避免把地点知识长期硬编码在 `visualization.py`
- [ ] 增强 `epoch_ai_gpu` 和相关算力采集器的源页面解析：即使公开 API 不给坐标，也继续尝试从详情页、HTML、内嵌 JSON、schema.org、OpenGraph、脚本变量和 PDF/新闻稿链接里抽地点线索
- [ ] 为未知位置算力中心增加外部富化策略评估：可选接入公开知识源或搜索兜底，只抓“站点名/园区名/城市名”级别线索，不直接抓经纬度结论，并把结果作为候选证据而不是真值
- [ ] 为算力中心建立 `operator / cluster name / facility alias` 归一化层，先解决 `xAI / Colossus / Memphis`、`OpenAI / Stargate`、`CoreWeave`、`Lambda`、`Crusoe` 这类同一对象多种写法导致的地点匹配失败
- [ ] 为估算位置增加更细的视觉和产品表达：除了问号角标，还要支持 tooltip/详情中的“估算依据”“精度级别”“最后核验时间”，并允许在设置中单独开关“仅看精确位置”
- [ ] 为国家级估算点设计更合理的落点策略：优先落在“该国主要算力/数据中心城市候选集”而不是几何质心，必要时同国多节点做稳定散列分配，避免大量节点堆在荒漠或海上
- [ ] 为未知位置算力中心建立人工校验工作流：支持导出待核验清单、记录人工确认结果，并把人工确认反哺到位置注册表，逐步减少问号点比例
+This file is the active backlog only. Completed history belongs in `docs/CHANGELOG.md`; detailed designs belong in `docs/plans/`.
+
+## Earth
+
+- [ ] Earth AI command entry: merge natural-language and speech-triggered LLM commands into the existing Earth search panel as described in [Agent Runtime, Earth LLM Command, And Speech Entry Plan](/home/ray/dev/linkong/planet/docs/plans/agents-earth-command-runtime-plan.md).
+- [ ] Earth action executor: implement safe visualization actions for layer toggles, batch highlights, filters, focus, result panels, and clear-highlight behavior.
+- [ ] Earth entity matching: support stable entity ids and batch matching for Beidou satellites, mainland China compute centers, BGP, news, vessels, and cables.
+- [x] High-precision country boundary tile framework: implement the static vector tile builder, versioned seed output, frontend bbox tile loader, debounce, in-flight dedupe, and LRU cache described in [Earth High Precision Boundary Tiles Plan](/home/ray/dev/linkong/planet/docs/plans/earth-high-precision-boundary-tiles-plan.md).
+- [x] Add the `pmtiles-mvt` frontend tile provider contract, MVT decoder dependencies, static PMTiles Nginx handling, collector artifact registration, production readiness check, and user operation docs for Earth boundaries.
+- [x] Split Earth boundary ingestion into standard source collectors (`earth_admin0_boundaries`, `earth_coastline`, `earth_claim_lines`) plus the downstream `earth_boundary_tiles` PMTiles builder.
+- [ ] Replace debug GeoJSON boundary tiles with the real `earth-boundaries-china-pov-v1.pmtiles` production artifact after audited admin-0 / coastline / claim-line sources and the PMTiles toolchain are available.
+- [ ] Import authoritative China POV / coastline / claim-line source packages through the three standard Earth boundary source collectors, then rebuild a versioned PMTiles artifact so highest zoom `8-10` preserves trusted source geometry instead of seed data.
+- [ ] Earth boundary data: acquire or generate auditable China POV geometry for Zangnan, Aksai Chin, Taiwan/Penghu, Diaoyu Dao and affiliated islands, Chiwei Yu, South China Sea islands, Kosovo, Gaza, and the official dashed maritime claim line before implementing final visual changes.
+- [ ] Earth high-resolution basemap tiles: implement the viewport-loaded imagery layer described in [Earth High Resolution Basemap Tiles Plan](/home/ray/dev/linkong/planet/docs/plans/earth-high-resolution-basemap-tiles-plan.md), using high-precision coastline as the alignment reference instead of replacing the globe with one huge texture.
+- [ ] Presentation controller ownership: replace the singleton card fallback in [presentation-controller.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/presentation-controller.js) with a presentation/card token check before BGP/News migrate onto the shared controller, so connectors only attach to their owning card.
+- [ ] BGP frontend maintainability: split [bgp.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp.js) by responsibility into data loading, marker rendering, overlays, and animation once the current interaction behavior is stable.
+- [ ] Optional BGP marker experiment: evaluate HTML markers for BGP incident/collector points if WebGL marker density or fixed screen-size clickability becomes a real blocker.
+- [ ] Earth news cruise: connect Earth news to the generic cruise queue via a news adapter rather than coupling news-specific sequencing into `main.js`.
+
+## Compute Centers And Location
+
+- [ ] Unknown compute-center locations: continue reducing unresolved records through the shared location pipeline, with confidence, precision, reason, and verification date preserved in GeoJSON/details.
+- [ ] Compute-center registry: keep expanding the local canonical location registry with `canonical_name`, aliases, operator, country/region/city, coordinates, confidence, and source notes.
+- [ ] Compute-center enrichment: improve source-page parsing for Epoch AI and related collectors by extracting location clues from detail pages, embedded JSON, schema.org, OpenGraph, script variables, PDFs, and press releases.
+- [ ] Compute-center identity normalization: normalize operator / cluster / facility aliases such as `xAI / Colossus / Memphis`, `OpenAI / Stargate`, `CoreWeave`, `Lambda`, and `Crusoe`.
+- [ ] Compute-center manual review: add an export/review/import workflow for unresolved or estimated locations and feed confirmed results back into the registry.
+
+## AIS / Vessels
+
+- [ ] AIS aggregation strategy v4: expose source priority, field-level merge rules, freshness windows, and protected dynamic-field rules in configuration, with validation and strategy version returned by vessel APIs.
+- [ ] AIS vessel enrichment v5: add asynchronous vessel profile enrichment for ship type detail, AIS class, flag, dimensions, build year, operator, and cached media. Do not fetch third-party pages in the realtime AIS request path.
+- [ ] AIS identity cleanup: continue identifying vessels whose display name is only `MMSI <number>` and backfill names from AISStream static messages, BarentsWatch static fields, or enrichment cache.
+
+## AI Provider And Agents
+
+- [ ] Agent runtime foundation: add auditable agent runs, steps, evidence, proposals, and the Agent operations UI described in [Agent Runtime, Earth LLM Command, And Speech Entry Plan](/home/ray/dev/linkong/planet/docs/plans/agents-earth-command-runtime-plan.md).
+- [ ] Agent tool protocol: add backend JSON tool-call fallback, optional provider-native tool compatibility, tool whitelist validation, and policy-gated proposal application.
+- [ ] Speech/ASR integration for agents: add provider-neutral transcription settings and API, defaulting to Whisper-compatible API providers while keeping text commands usable when ASR is unavailable.
+- [ ] Earth voice wake: add device-local configurable wake-word preferences, microphone fallback states, and post-wake instruction upload for Earth commands.
+- [ ] AI provider compatibility center: move provider/model compatibility rules into a JSON/YAML config read by runtime, instead of continuing to scatter provider-specific branches through Python code.
+- [ ] Provider compatibility coverage: add explicit config for OpenAI, Anthropic, MiniMax, Ollama, Moonshot, DeepSeek, Qwen, GLM, Gemini, OpenRouter, vLLM, LM Studio, and One API.
+- [ ] Compatibility schema: cover adapter type, base URL pattern, auth header, thinking/reasoning defaults, stream path, tool-call capability, multimodal capability, and provider-specific request patches.
+- [ ] BGP geography fallback: evaluate `inetnum` / `inet6num` whois as a finer fallback layer after `prefix_geography`, `OpenGeoFeed`, and RIR delegated data.
+
+## Platform
+
+- [ ] Earth preferences scope: keep current device-local Earth preferences in `localStorage`; only design backend user preferences if account-level synchronization becomes a real product requirement.
+- [ ] System logs: finish a usable Planet log viewing flow that covers backend, frontend, AI Provider, and collector/task logs, with filtering and tailing.
+- [ ] Console UI modernization: gradually replace Ant Design with Planet-owned components and a consistent Tabler Icons based icon system.
+- [ ] Earth live sync: design a unified realtime invalidation path for summary/BGP/satellite updates if polling and current WebSocket channels become insufficient.
+
+## Archive
+
+Archived items stay here so old context is not lost. Completed items remain checked; obsolete, invalid, or superseded items stay unchecked and include the reason.
+
+### Completed
+
+- [x] Refined BGP observer and anomaly `hover/click` feel.
+- [x] Added BGP anomaly relationship display with cables / regions.
+- [x] Added the Earth BGP activity layer so the map still feels alive when incident density is low.
+- [x] Added BGP state expression for stable observation, local fluctuation, and active incident states.
+- [x] Reframed "no active incident" as "observation network is running; no aggregate incident detected".
+- [x] Added collector / region recent activity scoring.
+- [x] Replaced oversized BGP incident glow with compact incident core plus outward pulse rings.
+- [x] Added BGP incident symbol types instead of using one generic bright marker.
+- [x] Switched BGP incident geography from collector-centric to prefix-centric priority.
+- [x] Added `prefix_geography` as a separate data layer instead of treating `prefix_scope` as prefix geography.
+- [x] Added IPtoASN / IPtoCountry as the main prefix-centric geography source.
+- [x] Added OpenGeoFeed as a high-quality prefix geography override source.
+- [x] Made RIR delegated data a prefix geography fallback rather than the primary source.
+- [x] Added route leak and path instability / flap detectors after the activity layer work.
+
+### Obsolete Or Superseded
+
+- [ ] AIS v3.1 old `/geo/vessels` full-merge requirement. Superseded by `/api/v1/vessels/snapshot`, controlled legacy fallback, and diagnostics in the AIS aggregation plan.
+- [ ] AIS v3.2 old framing of AISStream as a batch collector that needed conversion. Superseded by the implemented long-lived AISStream collector and realtime stream UI.
+- [ ] AIS v3.3 old one-shot REST progress semantics for AISStream. Superseded by realtime stream status handling.
+- [ ] AIS v3.4 broad identity cleanup wording. Folded into the active AIS identity cleanup and v5 enrichment tasks.
+- [ ] Earth surface material overlay for boundary calibration. Superseded by the high-precision boundary tile plan; future work must use source-faithful boundary/coastline data rather than overlay calibration against the coarse base map.
+- [ ] Hardcoded Earth news source extraction as a standalone task. Superseded by the broader Earth news source configuration and collector plans.
+- [ ] Country-level compute-center fallback placement as a standalone task. Superseded by the shared location pipeline and registry/manual-review backlog.
--- a/2
+++ b/2
@@ -1 +1 @@
-0.52.0
+0.60.0
--- a/aiprovider/Dockerfile
+++ b/aiprovider/Dockerfile
@@ -27,4 +27,7 @@ COPY aiprovider /app/aiprovider

 EXPOSE 8010

-CMD ["uv", "run", "--frozen", "--no-dev", "--project", "/app", "python", "-m", "uvicorn", "aiprovider.main:app", "--host", "0.0.0.0", "--port", "8010", "--reload"]
+HEALTHCHECK --interval=30s --timeout=5s --retries=3 \
+    CMD curl -fsS http://127.0.0.1:8010/health >/dev/null || exit 1
+
+CMD ["uv", "run", "--frozen", "--no-dev", "--project", "/app", "python", "-m", "uvicorn", "aiprovider.main:app", "--host", "0.0.0.0", "--port", "8010"]
--- a/aiprovider/config.py
+++ b/aiprovider/config.py
@@ -17,9 +17,6 @@ class Settings(BaseSettings):
    AI_HTTP_RETRY_ATTEMPTS: int = 2
    AI_MAX_TOKENS: int = 1200
    AI_ANTHROPIC_VERSION: str = "2023-06-01"
-    AI_ANALYSIS_SYSTEM_PROMPT: str = (
-        "你是态势感知分析助手。请基于输入的上下文、观测与约束，输出结构化、克制、可执行的分析。"
-    )

    AI_PROVIDER_SERVICE_TOKEN: str = ""

--- a/aiprovider/provider_service.py
+++ b/aiprovider/provider_service.py
@@ -62,7 +62,6 @@ class ProviderService:
        self.anthropic_version = str(
            overrides.get("anthropic_version") or settings.AI_ANTHROPIC_VERSION
        )
-        self.system_prompt = settings.AI_ANALYSIS_SYSTEM_PROMPT

    def get_status(self) -> AIProviderStatusResponse:
        enabled = self.provider != "disabled"
@@ -95,15 +94,20 @@ class ProviderService:
        prompt = self._build_prompt(payload)

        if self.provider_api == "openai-completions":
-            data = await self._request_openai_compatible(model, prompt)
+            data = await self._request_openai_compatible(model, prompt, payload.system_prompt)
            content = self._extract_openai_content(data)
            content_blocks = self._extract_openai_blocks(data)
        elif self.provider_api == "anthropic-messages":
-            data = await self._request_anthropic_messages(model, prompt, payload.thinking)
+            data = await self._request_anthropic_messages(
+                model,
+                prompt,
+                payload.thinking,
+                payload.system_prompt,
+            )
            content = self._extract_anthropic_content(data)
            content_blocks = self._extract_anthropic_blocks(data)
        elif self.provider_api == "ollama-generate":
-            data = await self._request_ollama(model, prompt)
+            data = await self._request_ollama(model, prompt, payload.system_prompt)
            content = self._extract_ollama_content(data)
            content_blocks = self._extract_ollama_blocks(data)
        else:
@@ -139,19 +143,28 @@ class ProviderService:
            sections.append("约束条件:\n" + "\n".join(f"- {item}" for item in payload.constraints))
        if payload.context:
            sections.append(f"附加上下文:\n{payload.context}")
-        sections.append(
-            "请输出: 1) 态势摘要 2) 关键风险 3) 研判依据 4) 建议动作 5) 还缺少的数据。"
-        )
        return "\n\n".join(sections)

-    async def _request_openai_compatible(self, model: str, prompt: str) -> dict[str, Any]:
+    def _resolve_system_prompt(self, system_prompt: str | None) -> str | None:
+        resolved = str(system_prompt or "").strip()
+        return resolved or None
+
+    async def _request_openai_compatible(
+        self,
+        model: str,
+        prompt: str,
+        system_prompt: str | None = None,
+    ) -> dict[str, Any]:
+        messages = []
+        resolved_system_prompt = self._resolve_system_prompt(system_prompt)
+        if resolved_system_prompt:
+            messages.append({"role": "system", "content": resolved_system_prompt})
+        messages.append({"role": "user", "content": prompt})
        request_body = {
            "model": model,
-            "messages": [
-                {"role": "system", "content": self.system_prompt},
-                {"role": "user", "content": prompt},
-            ],
+            "messages": messages,
            "temperature": 0.2,
+            "max_tokens": self.max_tokens,
        }
        return await self._post(
            path="/chat/completions",
@@ -167,10 +180,10 @@ class ProviderService:
        model: str,
        prompt: str,
        thinking: dict[str, Any] | None = None,
+        system_prompt: str | None = None,
    ) -> dict[str, Any]:
        request_body = {
            "model": model,
-            "system": self.system_prompt,
            "messages": [
                {
                    "role": "user",
@@ -185,6 +198,9 @@ class ProviderService:
            "max_tokens": self.max_tokens,
            "temperature": 0.2,
        }
+        resolved_system_prompt = self._resolve_system_prompt(system_prompt)
+        if resolved_system_prompt:
+            request_body["system"] = resolved_system_prompt
        resolved_thinking = self._resolve_anthropic_thinking(thinking)
        if resolved_thinking:
            request_body["thinking"] = resolved_thinking
@@ -218,19 +234,28 @@ class ProviderService:
        model: str,
        prompt: str,
        thinking: dict[str, Any] | None = None,
+        system_prompt: str | None = None,
    ) -> dict[str, Any]:
-        return await self._request_anthropic_messages(model, prompt, thinking)
+        return await self._request_anthropic_messages(model, prompt, thinking, system_prompt)

-    async def _request_ollama(self, model: str, prompt: str) -> dict[str, Any]:
+    async def _request_ollama(
+        self,
+        model: str,
+        prompt: str,
+        system_prompt: str | None = None,
+    ) -> dict[str, Any]:
        request_body = {
            "model": model,
            "stream": False,
-            "system": self.system_prompt,
            "prompt": prompt,
            "options": {
                "temperature": 0.2,
+                "num_predict": self.max_tokens,
            },
        }
+        resolved_system_prompt = self._resolve_system_prompt(system_prompt)
+        if resolved_system_prompt:
+            request_body["system"] = resolved_system_prompt
        return await self._post(
            path="/api/generate",
            headers={
--- a/aiprovider/schemas.py
+++ b/aiprovider/schemas.py
@@ -13,10 +13,11 @@ class AIContentBlock(BaseModel):

 class SituationalAnalysisRequest(BaseModel):
    title: str = Field(..., min_length=1, max_length=200)
-    objective: str = Field(..., min_length=1, max_length=1000)
+    objective: str = Field(..., min_length=1, max_length=20000)
    context: dict[str, Any] = Field(default_factory=dict)
    observations: list[str] = Field(default_factory=list)
    constraints: list[str] = Field(default_factory=list)
+    system_prompt: str | None = Field(default=None, max_length=8000)
    preferred_model: str | None = Field(default=None, max_length=200)
    thinking: dict[str, Any] | None = None

--- a/backend/Dockerfile
+++ b/backend/Dockerfile
@@ -12,6 +12,7 @@ ENV PYTHONDONTWRITEBYTECODE=1
 ENV PYTHONUNBUFFERED=1
 ENV UV_COMPILE_BYTECODE=1
 ENV UV_LINK_MODE=copy
+ENV PYTHONPATH=/app/backend

 RUN apt-get update && apt-get install -y --no-install-recommends \
    curl \
@@ -25,4 +26,7 @@ COPY VERSION /app/VERSION

 EXPOSE 8000

-CMD ["uv", "run", "--frozen", "--no-dev", "--project", "/app", "python", "-m", "uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]
+HEALTHCHECK --interval=30s --timeout=5s --retries=3 \
+    CMD curl -fsS http://127.0.0.1:8000/health >/dev/null || exit 1
+
+CMD ["uv", "run", "--frozen", "--no-dev", "--project", "/app", "python", "-m", "uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000"]
--- a/backend/app/ai_tasks/init.py
+++ b/backend/app/ai_tasks/init.py
@@ -0,0 +1,2 @@
+"""AI task prompt registry and runtime helpers."""
+
--- a/backend/app/ai_tasks/default_prompts.json
+++ b/backend/app/ai_tasks/default_prompts.json
@@ -0,0 +1,74 @@
+[
+  {
+    "key": "earth.news.enrich",
+    "label": "Earth 新闻汉化与定位",
+    "group": "Earth 新闻",
+    "version": "2026-05-16.2",
+    "system_prompt": "",
+    "prompt": "Return exactly one strict JSON object with a location object and a localizations object. Infer the most likely physical event location and produce a faithful Simplified Chinese title plus a one-sentence newswire-style Chinese summary based only on the supplied RSS headline, description, source, and date. The summary should read like a concise breaking-news lead, not a label, slogan, or keyword headline."
+  },
+  {
+    "key": "alerts.brief",
+    "label": "系统告警研判",
+    "group": "告警研判",
+    "version": "2026-05-16.1",
+    "system_prompt": "你是告警研判助手。请基于输入的告警事实、上下文与约束，输出结构化、克制、可执行的值班研判；明确区分事实、推断与建议，不要夸大证据不足的风险。",
+    "prompt": "基于当前告警总量、严重度、状态、数据源分布与最近告警摘录，生成一份面向值班人员的简明告警态势简报，突出待处理风险、告警集中点和优先动作。"
+  },
+  {
+    "key": "alerts.situational.brief",
+    "label": "跨模块态势告警研判",
+    "group": "告警研判",
+    "version": "2026-05-16.1",
+    "system_prompt": "你是告警研判助手。请基于输入的告警事实、上下文与约束，输出结构化、克制、可执行的值班研判；明确区分事实、推断与建议，不要夸大证据不足的风险。",
+    "prompt": "综合系统告警、BGP incidents、BGP anomalies 与近期 BGP AI 简报，生成一份面向值班人员的态势告警简报，指出当前最需要关注的风险域、跨模块联动迹象和优先动作。"
+  },
+  {
+    "key": "bgp.brief",
+    "label": "BGP 态势简报",
+    "group": "BGP",
+    "version": "2026-05-16.2",
+    "system_prompt": "你是 BGP 值班分析师。请直接输出面向值班人员的中文 Markdown 简报，只写最终研判内容；不要复述用户需求、提示词、写作计划、字段清单或“我将如何回答”。",
+    "prompt": "基于当前 BGP incidents、anomalies、原始观测事件、观测站覆盖与 prefix geography 证据，生成一份面向操作员的简明态势简报，突出区域热点、观测偏差、当前风险、证据和优先动作。"
+  },
+  {
+    "key": "location.factcheck.normalize",
+    "label": "位置事实核查结构化",
+    "group": "位置解析",
+    "version": "2026-05-16.1",
+    "system_prompt": "",
+    "prompt": "Convert the supplied location factcheck text into exactly one strict JSON object. Extract only facts present in the text or original query."
+  },
+  {
+    "key": "location.factcheck.resolve",
+    "label": "位置事实核查兜底",
+    "group": "位置解析",
+    "version": "2026-05-16.1",
+    "system_prompt": "",
+    "prompt": "Return exactly one JSON object for the most likely physical location. Use only fact-checkable public knowledge; return null fields rather than guessing when evidence is weak."
+  },
+  {
+    "key": "datasource.mapping",
+    "label": "数据源映射生成",
+    "group": "采集配置",
+    "version": "2026-05-16.1",
+    "system_prompt": "",
+    "prompt": "Return only JSON for a deterministic mapping DSL. The JSON must contain source.items_path and fields. Do not include prose or code."
+  },
+  {
+    "key": "credential.guide",
+    "label": "采集器凭据教程",
+    "group": "采集配置",
+    "version": "2026-05-16.1",
+    "system_prompt": "",
+    "prompt": "生成一份中文采集器凭据配置教程。只能根据 context.search_evidence 中的来源生成教程；如果证据不足，明确说明需要以官方页面为准。"
+  },
+  {
+    "key": "ai.connection_test",
+    "label": "AI Provider 连接测试",
+    "group": "运维测试",
+    "version": "2026-05-16.1",
+    "system_prompt": "",
+    "prompt": "Reply OK."
+  }
+]
--- a/backend/app/ai_tasks/prompts.py
+++ b/backend/app/ai_tasks/prompts.py
@@ -0,0 +1,182 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import UTC, datetime
+import json
+from functools import lru_cache
+from pathlib import Path
+from typing import Any
+
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models.system_setting import SystemSetting
+
+AI_PROMPTS_CATEGORY = "ai_prompts"
+DEFAULT_PROMPTS_PATH = Path(__file__).with_name("default_prompts.json")
+
+
+@dataclass(frozen=True)
+class AIPromptDefinition:
+    key: str
+    label: str
+    group: str
+    version: str
+    system_prompt: str
+    prompt: str
+
+
+@dataclass(frozen=True)
+class EffectiveAIPrompt:
+    key: str
+    label: str
+    group: str
+    version: str
+    default_system_prompt: str
+    default_prompt: str
+    system_prompt: str
+    prompt: str
+    is_custom: bool
+    updated_at: str | None = None
+
+
+@lru_cache(maxsize=1)
+def list_prompt_definitions() -> tuple[AIPromptDefinition, ...]:
+    raw_items = json.loads(DEFAULT_PROMPTS_PATH.read_text(encoding="utf-8"))
+    return tuple(
+        AIPromptDefinition(
+            key=str(item["key"]),
+            label=str(item["label"]),
+            group=str(item["group"]),
+            version=str(item["version"]),
+            system_prompt=str(item.get("system_prompt") or ""),
+            prompt=str(item.get("prompt") or ""),
+        )
+        for item in raw_items
+    )
+
+
+def get_prompt_definition(task_key: str) -> AIPromptDefinition:
+    for definition in list_prompt_definitions():
+        if definition.key == task_key:
+            return definition
+    raise KeyError(task_key)
+
+
+async def _get_prompt_setting(db: AsyncSession) -> SystemSetting | None:
+    result = await db.execute(
+        select(SystemSetting).where(SystemSetting.category == AI_PROMPTS_CATEGORY)
+    )
+    return result.scalar_one_or_none()
+
+
+def _normalize_overrides(payload: dict[str, Any] | None) -> dict[str, dict[str, Any]]:
+    raw = (payload or {}).get("overrides")
+    if not isinstance(raw, dict):
+        return {}
+    return {
+        str(key): dict(value)
+        for key, value in raw.items()
+        if isinstance(value, dict)
+    }
+
+
+async def get_prompt_overrides(db: AsyncSession) -> dict[str, dict[str, Any]]:
+    if not hasattr(db, "execute"):
+        return {}
+    setting = await _get_prompt_setting(db)
+    return _normalize_overrides(setting.payload if setting else None)
+
+
+def _effective_prompt(
+    definition: AIPromptDefinition,
+    override: dict[str, Any] | None,
+) -> EffectiveAIPrompt:
+    override = override or {}
+    custom_system = override.get("system_prompt")
+    custom_prompt = override.get("prompt")
+    has_custom_system = isinstance(custom_system, str)
+    has_custom_prompt = isinstance(custom_prompt, str)
+    return EffectiveAIPrompt(
+        key=definition.key,
+        label=definition.label,
+        group=definition.group,
+        version=definition.version,
+        default_system_prompt=definition.system_prompt,
+        default_prompt=definition.prompt,
+        system_prompt=custom_system if has_custom_system else definition.system_prompt,
+        prompt=custom_prompt if has_custom_prompt else definition.prompt,
+        is_custom=has_custom_system or has_custom_prompt,
+        updated_at=str(override.get("updated_at") or "") or None,
+    )
+
+
+async def list_effective_prompts(db: AsyncSession) -> list[EffectiveAIPrompt]:
+    overrides = await get_prompt_overrides(db)
+    return [
+        _effective_prompt(definition, overrides.get(definition.key))
+        for definition in list_prompt_definitions()
+    ]
+
+
+async def get_effective_prompt(db: AsyncSession | None, task_key: str) -> EffectiveAIPrompt:
+    definition = get_prompt_definition(task_key)
+    if db is None:
+        return _effective_prompt(definition, None)
+    overrides = await get_prompt_overrides(db)
+    return _effective_prompt(definition, overrides.get(task_key))
+
+
+async def save_prompt_override(
+    db: AsyncSession,
+    task_key: str,
+    *,
+    system_prompt: str,
+    prompt: str,
+) -> EffectiveAIPrompt:
+    definition = get_prompt_definition(task_key)
+    setting = await _get_prompt_setting(db)
+    payload = dict(setting.payload or {}) if setting else {}
+    overrides = _normalize_overrides(payload)
+    overrides[definition.key] = {
+        "system_prompt": system_prompt,
+        "prompt": prompt,
+        "updated_at": datetime.now(UTC).isoformat().replace("+00:00", "Z"),
+    }
+    payload["overrides"] = overrides
+    if setting is None:
+        setting = SystemSetting(category=AI_PROMPTS_CATEGORY, payload=payload)
+        db.add(setting)
+    else:
+        setting.payload = payload
+    await db.commit()
+    return _effective_prompt(definition, overrides[definition.key])
+
+
+async def reset_prompt_override(db: AsyncSession, task_key: str) -> EffectiveAIPrompt:
+    definition = get_prompt_definition(task_key)
+    setting = await _get_prompt_setting(db)
+    if setting is None:
+        return _effective_prompt(definition, None)
+    payload = dict(setting.payload or {})
+    overrides = _normalize_overrides(payload)
+    overrides.pop(definition.key, None)
+    payload["overrides"] = overrides
+    setting.payload = payload
+    await db.commit()
+    return _effective_prompt(definition, None)
+
+
+def serialize_effective_prompt(prompt: EffectiveAIPrompt) -> dict[str, Any]:
+    return {
+        "key": prompt.key,
+        "label": prompt.label,
+        "group": prompt.group,
+        "version": prompt.version,
+        "default_system_prompt": prompt.default_system_prompt,
+        "default_prompt": prompt.default_prompt,
+        "system_prompt": prompt.system_prompt,
+        "prompt": prompt.prompt,
+        "is_custom": prompt.is_custom,
+        "updated_at": prompt.updated_at,
+    }
--- a/backend/app/api/main.py
+++ b/backend/app/api/main.py
@@ -6,6 +6,7 @@ from app.api.v1 import (
    datasource_config,
    datasources,
    docs,
+    earth,
    tasks,
    dashboard,
    alerts,
@@ -18,6 +19,7 @@ from app.api.v1 import (
    vessels,
    bgp,
    news,
+    realtime_sources,
    system_control,
    tv,
 )
@@ -33,6 +35,7 @@ api_router.include_router(
 api_router.include_router(datasources.router, prefix="/datasources", tags=["datasources"])
 api_router.include_router(collected_data.router, prefix="/collected", tags=["collected-data"])
 api_router.include_router(docs.router, prefix="/docs", tags=["docs"])
+api_router.include_router(earth.router, prefix="/earth", tags=["earth"])
 api_router.include_router(tasks.router, prefix="/tasks", tags=["tasks"])
 api_router.include_router(dashboard.router, prefix="/dashboard", tags=["dashboard"])
 api_router.include_router(alerts.router, prefix="/alerts", tags=["alerts"])
@@ -50,3 +53,4 @@ api_router.include_router(vessels.router, prefix="/vessels", tags=["vessels"])
 api_router.include_router(bgp.router, prefix="/bgp", tags=["bgp"])
 api_router.include_router(tv.router, prefix="/tv", tags=["tv"])
 api_router.include_router(news.router, prefix="/news", tags=["news"])
+api_router.include_router(realtime_sources.router, prefix="/realtime-sources", tags=["realtime-sources"])
--- a/backend/app/api/v1/bgp.py
+++ b/backend/app/api/v1/bgp.py
@@ -341,6 +341,7 @@ async def collect_bgp_collector_location(
                provider_client=provider_client,
                query=query,
                entity_type="bgp_collector",
+                db=db,
                attempted_queries=attempted_queries,
                search_evidence=search_result.evidence,
            )
--- a/backend/app/api/v1/collected_data.py
+++ b/backend/app/api/v1/collected_data.py
@@ -102,7 +102,12 @@ def build_search_rank_sql(search: Optional[str]) -> str:
    """


-def serialize_collected_row(row, source_name_map: dict[str, str] | None = None) -> dict:
+def serialize_collected_row(
+    row,
+    source_name_map: dict[str, str] | None = None,
+    *,
+    include_metadata: bool = True,
+) -> dict:
    metadata = row[7]
    source = row[1]
    return {
@@ -120,7 +125,7 @@ def serialize_collected_row(row, source_name_map: dict[str, str] | None = None)
        "longitude": get_metadata_field(metadata, "longitude"),
        "value": get_metadata_field(metadata, "value"),
        "unit": get_metadata_field(metadata, "unit"),
-        "metadata": metadata,
+        "metadata": metadata if include_metadata else None,
        "cores": get_metadata_field(metadata, "cores"),
        "rmax": get_metadata_field(metadata, "rmax"),
        "rpeak": get_metadata_field(metadata, "rpeak"),
@@ -145,6 +150,7 @@ async def list_collected_data(
    search: Optional[str] = Query(None, description="搜索名称"),
    page: int = Query(1, ge=1, description="页码"),
    page_size: int = Query(20, ge=1, le=100, description="每页数量"),
+    include_metadata: bool = Query(True, description="是否返回完整 metadata 字段"),
    current_user: User = Depends(get_current_user),
    db: AsyncSession = Depends(get_db),
 ):
@@ -201,7 +207,7 @@ async def list_collected_data(

    data = []
    for row in rows:
-        data.append(serialize_collected_row(row[:11], source_name_map))
+        data.append(serialize_collected_row(row[:11], source_name_map, include_metadata=include_metadata))

    return {
        "total": total,
--- a/backend/app/api/v1/datasource_config.py
+++ b/backend/app/api/v1/datasource_config.py
@@ -22,6 +22,7 @@ from app.models.vessel import AISRawObservation, AISSourceHealth
 from app.core.security import get_current_user
 from app.core.cache import cache
 from app.core.time import to_iso8601_utc
+from app.ai_tasks.prompts import get_effective_prompt
 from app.schemas.ai import SituationalAnalysisRequest
 from app.services.ai_client import AIProviderClient, get_ai_provider_client
 from app.services.datasource_mapping import (
@@ -41,6 +42,8 @@ from app.services.custom_datasource_runtime import (
    stop_custom_stream,
    test_websocket_config,
 )
+
+DATASOURCE_MAPPING_PROMPT_KEY = "datasource.mapping"
 from app.services.datasource_connectivity import (
    get_builtin_connection_status,
    save_connectivity_success,
@@ -50,6 +53,15 @@ from app.services.datasource_connectivity import (

 router = APIRouter()

+def _default_builtin_config(name: str) -> dict[str, Any]:
+    return {"timeout": 30, "retry": 3}
+
+
+def _default_builtin_source_type(name: str) -> str:
+    if name == "aisstream_vessels":
+        return "websocket"
+    return "http"
+

 class DataSourceConfigCreate(BaseModel):
    name: str = Field(..., min_length=1, max_length=100)
@@ -364,7 +376,7 @@ async def list_all_datasources(
    db: AsyncSession = Depends(get_db),
 ):
    """List all data sources: YAML defaults + DB overrides"""
-    from app.core.data_sources import COLLECTOR_URL_KEYS, get_data_sources_config
+    from app.core.data_sources import get_data_sources_config

    config = get_data_sources_config()

@@ -372,20 +384,22 @@ async def list_all_datasources(
    db_configs = {c.name: c for c in db_query.scalars().all()}

    result = []
-    for name, yaml_key in COLLECTOR_URL_KEYS.items():
+    for name, metadata in DEFAULT_DATASOURCES.items():
        yaml_url = config.get_yaml_url(name)
        db_config = db_configs.get(name)
+        default_config = _default_builtin_config(name)
+        default_url = yaml_url

        result.append(
            {
                "name": name,
-                "default_url": yaml_url,
-                "endpoint": db_config.endpoint if db_config else yaml_url,
+                "default_url": default_url,
+                "endpoint": db_config.endpoint if db_config else default_url,
                "is_overridden": db_config is not None and db_config.endpoint != yaml_url
-                if yaml_url
+                if default_url
                else db_config is not None,
                "is_active": db_config.is_active if db_config else True,
-                "source_type": db_config.source_type if db_config else "http",
+                "source_type": db_config.source_type if db_config else _default_builtin_source_type(name),
                "auth_type": db_config.auth_type if db_config else "none",
                "auth_configured": {
                    "api_key": bool((db_config.auth_config or {}).get("api_key"))
@@ -393,11 +407,11 @@ async def list_all_datasources(
                    else False,
                },
                "headers": db_config.headers if db_config else {},
-                "config": strip_connectivity_validation(db_config.config if db_config else {}),
+                "config": strip_connectivity_validation(db_config.config if db_config else default_config),
                "config_id": db_config.id if db_config else None,
                "description": db_config.description
                if db_config
-                else f"Data source from YAML: {yaml_key}",
+                else f"内置采集器默认配置：{metadata.get('display_name') or metadata.get('name') or name}",
            }
        )

@@ -757,14 +771,12 @@ async def propose_datasource_mapping(
    generated_by = "heuristic"
    if payload.use_ai:
        try:
+            prompt = await get_effective_prompt(db, DATASOURCE_MAPPING_PROMPT_KEY)
            response = await ai_client.analyze(
                SituationalAnalysisRequest(
                    title=f"Generate datasource mapping for {schema.key}",
-                    objective=(
-                        "Return only JSON for a deterministic mapping DSL. "
-                        "The JSON must contain source.items_path and fields. "
-                        "Do not include prose or code."
-                    ),
+                    objective=prompt.prompt,
+                    system_prompt=prompt.system_prompt or None,
                    context={
                        "target_schema": schema.to_dict(),
                        "sample_payload": redacted_sample,
--- a/backend/app/api/v1/datasources.py
+++ b/backend/app/api/v1/datasources.py
@@ -18,6 +18,8 @@ from app.models.datasource import DataSource
 from app.models.datasource_config import DataSourceConfig
 from app.models.task import CollectionTask
 from app.models.user import User
+from app.models.vessel import AISRawObservation
+from app.services.vessel_ais_aggregation import VESSEL_AIS_SCHEMA
 from app.services.scheduler import (
    cancel_running_collector_now,
    get_latest_task_id_for_datasource,
@@ -148,6 +150,33 @@ async def _load_latest_task_ids(
    return {datasource_id: task_id for datasource_id, task_id in result.all()}


+async def _load_latest_tasks(
+    db: AsyncSession,
+    datasource_ids: list[int],
+) -> dict[int, CollectionTask]:
+    if not datasource_ids:
+        return {}
+
+    ranked_tasks = (
+        select(
+            CollectionTask.id.label("task_id"),
+            CollectionTask.datasource_id.label("datasource_id"),
+            func.row_number().over(
+                partition_by=CollectionTask.datasource_id,
+                order_by=(CollectionTask.created_at.desc().nullslast(), CollectionTask.id.desc()),
+            ).label("row_num"),
+        )
+        .where(CollectionTask.datasource_id.in_(datasource_ids))
+        .subquery()
+    )
+    result = await db.execute(
+        select(CollectionTask)
+        .join(ranked_tasks, CollectionTask.id == ranked_tasks.c.task_id)
+        .where(ranked_tasks.c.row_num == 1)
+    )
+    return {task.datasource_id: task for task in result.scalars().all()}
+
+
 async def _load_collected_record_counts(
    db: AsyncSession,
    sources: list[str],
@@ -161,7 +190,26 @@ async def _load_collected_record_counts(
        .where(CollectedData.is_current.is_(True))
        .group_by(CollectedData.source)
    )
-    return {source: int(count or 0) for source, count in result.all()}
+    counts = {source: int(count or 0) for source, count in result.all()}
+
+    vessel_sources = [
+        source
+        for source in sources
+        if datasource_metadata(source)["credential_provider"] in {"aisstream", "barentswatch"}
+        or "vessel" in source
+        or "ais" in source
+    ]
+    if vessel_sources:
+        raw_result = await db.execute(
+            select(AISRawObservation.source, func.count(AISRawObservation.id))
+            .where(AISRawObservation.target_schema == VESSEL_AIS_SCHEMA)
+            .where(AISRawObservation.source.in_(vessel_sources))
+            .group_by(AISRawObservation.source)
+        )
+        for source, count in raw_result.all():
+            counts[source] = max(counts.get(source, 0), int(count or 0))
+
+    return counts


 async def _load_datasource_endpoint_overrides(
@@ -187,7 +235,7 @@ async def _load_datasource_endpoint_overrides(
 async def _load_datasource_list_context(
    db: AsyncSession,
    datasources: list[DataSource],
-) -> tuple[dict[int, CollectionTask], dict[str, str]]:
+) -> tuple[dict[int, CollectionTask], dict[int, CollectionTask], dict[str, str]]:
    datasource_ids = [datasource.id for datasource in datasources]
    sources = [datasource.source for datasource in datasources]

@@ -211,8 +259,9 @@ async def _load_datasource_list_context(
    if stale_datasource_ids:
        running_tasks = await _load_latest_running_tasks(db, datasource_ids)

+    latest_tasks = await _load_latest_tasks(db, datasource_ids)
    endpoint_overrides = await _load_datasource_endpoint_overrides(db, sources)
-    return running_tasks, endpoint_overrides
+    return running_tasks, latest_tasks, endpoint_overrides


 def _apply_datasource_query_filters(
@@ -230,11 +279,6 @@ def _apply_datasource_query_filters(
        query = query.where(DataSource.is_active == is_active)
    if priority:
        query = query.where(DataSource.priority == priority)
-    if run_status and run_status not in {"running", "collected", "uncollected"}:
-        if run_status == "not_run":
-            query = query.where(DataSource.last_status.is_(None))
-        else:
-            query = query.where(DataSource.last_status == run_status)
    if q:
        like_value = f"%{q.strip()}%"
        query = query.where(
@@ -251,15 +295,25 @@ def _filter_datasources_in_memory(
    datasources: list[DataSource],
    *,
    running_tasks: dict[int, CollectionTask],
+    latest_tasks: dict[int, CollectionTask] | None = None,
    record_counts: dict[str, int],
    product: Optional[str] = None,
    run_status: Optional[str] = None,
    collected: Optional[bool] = None,
    credential_status: Optional[str] = None,
 ) -> list[DataSource]:
+    latest_tasks = latest_tasks or {}
    filtered: list[DataSource] = []
    for datasource in datasources:
        record_count = record_counts.get(datasource.source, 0)
+        latest_task = latest_tasks.get(datasource.id)
+        effective_status = (
+            "running"
+            if datasource.id in running_tasks
+            else latest_task.status
+            if latest_task is not None
+            else datasource.last_status
+        )
        if product and datasource_product_key(datasource) != product:
            continue
        if collected is not None and (record_count > 0) != collected:
@@ -270,6 +324,10 @@ def _filter_datasources_in_memory(
                continue
        if run_status == "running" and datasource.id not in running_tasks:
            continue
+        if run_status == "not_run" and effective_status is not None:
+            continue
+        if run_status not in {None, "running", "not_run", "collected", "uncollected"} and effective_status != run_status:
+            continue
        if run_status == "collected" and record_count <= 0:
            continue
        if run_status == "uncollected" and record_count > 0:
@@ -618,11 +676,12 @@ async def list_datasources(

    collector_list = []
    config = get_data_sources_config()
-    running_tasks, endpoint_overrides = await _load_datasource_list_context(db, datasources)
+    running_tasks, latest_tasks, endpoint_overrides = await _load_datasource_list_context(db, datasources)
    record_counts = await _load_collected_record_counts(db, [datasource.source for datasource in datasources])
    datasources = _filter_datasources_in_memory(
        datasources,
        running_tasks=running_tasks,
+        latest_tasks=latest_tasks,
        record_counts=record_counts,
        product=product,
        run_status=run_status,
@@ -631,9 +690,11 @@ async def list_datasources(
    )
    for datasource in datasources:
        running_task = running_tasks.get(datasource.id)
+        latest_task = latest_tasks.get(datasource.id)
+        display_task = running_task or latest_task
        endpoint = endpoint_overrides.get(datasource.source) or config.get_yaml_url(datasource.source)
-        last_run_at = datasource.last_run_at
-        last_status = datasource.last_status
+        last_run_at = datasource.last_run_at or (latest_task.completed_at if latest_task else None)
+        last_status = datasource.last_status or (latest_task.status if latest_task else None)
        collected_records = record_counts.get(datasource.source, 0)

        collector_list.append(
@@ -654,16 +715,17 @@ async def list_datasources(
                "last_run_at": to_iso8601_utc(last_run_at),
                "last_status": last_status,
                "is_running": running_task is not None,
-                "task_id": running_task.id if running_task else None,
-                "progress": running_task.progress if running_task else None,
-                "phase": running_task.phase if running_task else None,
-                "phase_progress": running_task.phase_progress if running_task else None,
-                "phase_message": running_task.phase_message if running_task else None,
-                "phase_current": running_task.phase_current if running_task else None,
-                "phase_total": running_task.phase_total if running_task else None,
-                "phase_unit": running_task.phase_unit if running_task else None,
-                "records_processed": running_task.records_processed if running_task else None,
-                "total_records": running_task.total_records if running_task else None,
+                "task_id": display_task.id if display_task else None,
+                "progress": display_task.progress if display_task else None,
+                "phase": display_task.phase if display_task else None,
+                "phase_progress": display_task.phase_progress if display_task else None,
+                "phase_message": display_task.phase_message if display_task else None,
+                "phase_current": display_task.phase_current if display_task else None,
+                "phase_total": display_task.phase_total if display_task else None,
+                "phase_unit": display_task.phase_unit if display_task else None,
+                "records_processed": display_task.records_processed if display_task else None,
+                "total_records": display_task.total_records if display_task else None,
+                "error_message": display_task.error_message if display_task else None,
                "collected_records": collected_records,
                "has_collected_data": collected_records > 0,
            }
@@ -708,11 +770,12 @@ async def trigger_datasource_batch(

    result = await db.execute(query)
    datasources = result.scalars().all()
-    running_tasks, _ = await _load_datasource_list_context(db, datasources)
+    running_tasks, latest_tasks, _ = await _load_datasource_list_context(db, datasources)
    record_counts = await _load_collected_record_counts(db, [datasource.source for datasource in datasources])
    datasources = _filter_datasources_in_memory(
        datasources,
        running_tasks=running_tasks,
+        latest_tasks=latest_tasks,
        record_counts=record_counts,
        product=None if payload.source_ids else payload.product,
        run_status=None if payload.source_ids else payload.run_status,
@@ -914,6 +977,14 @@ async def get_task_status(
            raise HTTPException(status_code=404, detail="Task not found")
    else:
        task = await get_running_task(db, datasource.id)
+        if task is None:
+            result = await db.execute(
+                select(CollectionTask)
+                .where(CollectionTask.datasource_id == datasource.id)
+                .order_by(CollectionTask.created_at.desc().nullslast(), CollectionTask.id.desc())
+                .limit(1)
+            )
+            task = result.scalar_one_or_none()

    if not task:
        return {
@@ -942,4 +1013,5 @@ async def get_task_status(
        "records_processed": task.records_processed,
        "total_records": task.total_records,
        "status": task.status,
+        "error_message": task.error_message,
    }
--- a/backend/app/api/v1/earth.py
+++ b/backend/app/api/v1/earth.py
@@ -0,0 +1,247 @@
+"""Earth asset management APIs."""
+
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Any
+from uuid import uuid4
+
+from fastapi import APIRouter, Depends, File, HTTPException, Request, UploadFile, status
+from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer
+from pydantic import BaseModel, Field
+from sqlalchemy import delete, select, text
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.core.security import decode_token, get_current_user, redis_client
+from app.db.session import get_db
+from app.models.system_setting import SystemSetting
+from app.models.user import User
+from app.services.earth_boundaries import (
+    EarthBoundaryBuildError,
+    get_boundary_build_status,
+    get_boundary_status,
+    save_boundary_config,
+    start_boundary_build_job,
+)
+
+
+router = APIRouter()
+optional_bearer = HTTPBearer(auto_error=False)
+REPO_ROOT = Path(__file__).resolve().parents[4]
+EARTH_BRAND_ASSET_DIR = REPO_ROOT / "data" / "earth-brand"
+EARTH_BRAND_ASSET_URL_PREFIX = "/earth-brand-assets"
+EARTH_BRAND_CATEGORY = "earth_brand"
+MAX_EARTH_BRAND_ASSET_BYTES = 3 * 1024 * 1024
+ALLOWED_EARTH_BRAND_EXTENSIONS = {".png", ".jpg", ".jpeg", ".webp", ".svg"}
+
+DEFAULT_EARTH_BRAND = {
+    "logo_src": "/earth/assets/brand/earth-logo.png",
+    "title_src": "/earth/assets/brand/title-zh.png",
+    "title_text": "智能星球计划",
+    "subtitle": "现实层宇宙全息感知系统",
+    "description": "卫星 · 海底光缆 · 算力基础设施",
+    "aria_label": "智能星球计划品牌标识",
+    "title_alt": "智能星球计划",
+}
+
+
+class EarthBoundaryConfigPayload(BaseModel):
+    config: dict[str, Any] = Field(default_factory=dict)
+
+
+class EarthBrandPayload(BaseModel):
+    logo_src: str = Field(default=DEFAULT_EARTH_BRAND["logo_src"], max_length=1000)
+    title_src: str = Field(default=DEFAULT_EARTH_BRAND["title_src"], max_length=1000)
+    title_text: str = Field(default=DEFAULT_EARTH_BRAND["title_text"], max_length=120)
+    subtitle: str = Field(default=DEFAULT_EARTH_BRAND["subtitle"], max_length=160)
+    description: str = Field(default=DEFAULT_EARTH_BRAND["description"], max_length=200)
+    aria_label: str = Field(default=DEFAULT_EARTH_BRAND["aria_label"], max_length=200)
+    title_alt: str = Field(default=DEFAULT_EARTH_BRAND["title_alt"], max_length=200)
+
+
+def _normalize_earth_brand_payload(payload: dict[str, Any] | None) -> dict[str, str]:
+    merged = DEFAULT_EARTH_BRAND.copy()
+    if payload:
+        for key in DEFAULT_EARTH_BRAND:
+            value = payload.get(key)
+            if value is not None:
+                merged[key] = str(value).strip()
+
+    if not merged["title_text"]:
+        merged["title_text"] = DEFAULT_EARTH_BRAND["title_text"]
+    if not merged["aria_label"]:
+        merged["aria_label"] = merged["title_text"]
+    if not merged["title_alt"]:
+        merged["title_alt"] = merged["title_text"]
+    return merged
+
+
+async def _get_earth_brand_record(db: AsyncSession) -> SystemSetting | None:
+    result = await db.execute(
+        select(SystemSetting).where(SystemSetting.category == EARTH_BRAND_CATEGORY)
+    )
+    return result.scalar_one_or_none()
+
+
+async def _get_earth_brand_payload(db: AsyncSession) -> dict[str, Any]:
+    record = await _get_earth_brand_record(db)
+    return {
+        "brand": _normalize_earth_brand_payload(record.payload if record else None),
+        "is_default": record is None,
+    }
+
+
+@router.get("/brand")
+async def get_earth_brand(db: AsyncSession = Depends(get_db)):
+    return await _get_earth_brand_payload(db)
+
+
+@router.put("/brand")
+async def update_earth_brand(
+    payload: EarthBrandPayload,
+    _current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    normalized = _normalize_earth_brand_payload(payload.model_dump())
+    record = await _get_earth_brand_record(db)
+    if record is None:
+        record = SystemSetting(category=EARTH_BRAND_CATEGORY, payload=normalized)
+        db.add(record)
+    else:
+        record.payload = normalized
+    await db.commit()
+    await db.refresh(record)
+    return {"status": "updated", "brand": _normalize_earth_brand_payload(record.payload), "is_default": False}
+
+
+@router.delete("/brand")
+@router.post("/brand/reset")
+async def reset_earth_brand(
+    _current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    await db.execute(delete(SystemSetting).where(SystemSetting.category == EARTH_BRAND_CATEGORY))
+    await db.commit()
+    return {"status": "reset", "brand": DEFAULT_EARTH_BRAND.copy(), "is_default": True}
+
+
+@router.post("/brand/assets")
+async def upload_earth_brand_asset(
+    file: UploadFile = File(...),
+    _current_user: User = Depends(get_current_user),
+):
+    original_name = file.filename or ""
+    extension = Path(original_name).suffix.lower()
+    if extension not in ALLOWED_EARTH_BRAND_EXTENSIONS:
+        raise HTTPException(
+            status_code=400,
+            detail={
+                "code": "unsupported_file_type",
+                "message": "Only png, jpg, jpeg, webp, and svg brand assets are supported.",
+            },
+        )
+
+    content = await file.read(MAX_EARTH_BRAND_ASSET_BYTES + 1)
+    if len(content) > MAX_EARTH_BRAND_ASSET_BYTES:
+        raise HTTPException(
+            status_code=400,
+            detail={
+                "code": "file_too_large",
+                "message": "Brand asset must be 3 MB or smaller.",
+            },
+        )
+
+    EARTH_BRAND_ASSET_DIR.mkdir(parents=True, exist_ok=True)
+    safe_name = f"{uuid4().hex}{extension}"
+    destination = EARTH_BRAND_ASSET_DIR / safe_name
+    destination.write_bytes(content)
+    asset_url = f"{EARTH_BRAND_ASSET_URL_PREFIX}/{safe_name}"
+    return {"url": asset_url, "filename": safe_name, "content_type": file.content_type}
+
+
+@router.get("/boundaries/status")
+async def get_earth_boundary_status():
+    return get_boundary_status()
+
+
+async def _get_optional_current_user(
+    credentials: HTTPAuthorizationCredentials | None = Depends(optional_bearer),
+    db: AsyncSession = Depends(get_db),
+) -> User | None:
+    if credentials is None:
+        return None
+    token = credentials.credentials
+    if redis_client.sismember("blacklisted_tokens", token):
+        return None
+    payload = decode_token(token)
+    if payload is None or payload.get("type") != "access":
+        return None
+    user_id = payload.get("sub")
+    if user_id is None:
+        return None
+    result = await db.execute(
+        text(
+            "SELECT id, username, email, password_hash, role, is_active, gatekeeper_groups FROM users WHERE id = :id"
+        ),
+        {"id": int(user_id)},
+    )
+    row = result.fetchone()
+    if row is None or not row[5]:
+        return None
+    user = User()
+    user.id = row[0]
+    user.username = row[1]
+    user.email = row[2]
+    user.password_hash = row[3]
+    user.role = row[4]
+    user.is_active = row[5]
+    user.gatekeeper_groups = row[6] or []
+    return user
+
+
+def _is_loopback_request(request: Request) -> bool:
+    host = request.client.host if request.client else ""
+    return host in {"127.0.0.1", "::1", "localhost"} or host.startswith("127.")
+
+
+def _require_local_or_user(request: Request, user: User | None) -> None:
+    if user is not None or _is_loopback_request(request):
+        return
+    raise HTTPException(
+        status_code=status.HTTP_401_UNAUTHORIZED,
+        detail="Authentication required outside localhost",
+    )
+
+
+@router.put("/boundaries/config")
+async def update_earth_boundary_config(
+    payload: EarthBoundaryConfigPayload,
+    _current_user: User = Depends(get_current_user),
+):
+    try:
+        return save_boundary_config(payload.config)
+    except EarthBoundaryBuildError as exc:
+        raise HTTPException(
+            status_code=400,
+            detail={"code": exc.code, "message": str(exc), "details": exc.details},
+        ) from exc
+
+
+@router.post("/boundaries/build")
+async def build_earth_boundary_assets(
+    request: Request,
+    current_user: User | None = Depends(_get_optional_current_user),
+):
+    _require_local_or_user(request, current_user)
+    try:
+        return await start_boundary_build_job()
+    except EarthBoundaryBuildError as exc:
+        raise HTTPException(
+            status_code=400,
+            detail={"code": exc.code, "message": str(exc), "details": exc.details},
+        ) from exc
+
+
+@router.get("/boundaries/build/status")
+async def get_earth_boundary_build_status():
+    return get_boundary_build_status()
--- a/backend/app/api/v1/layers.py
+++ b/backend/app/api/v1/layers.py
@@ -117,7 +117,7 @@ async def get_vessel_layer_snapshot(
        bbox=parsed_bbox,
        zoom=zoom,
        limit=limit,
-        vessel_type=vessel_type,
+        type_filter=vessel_type,
        since_minutes=since_minutes,
    )

--- a/backend/app/api/v1/news.py
+++ b/backend/app/api/v1/news.py
@@ -1,5 +1,7 @@
-from fastapi import APIRouter, Query
+from fastapi import APIRouter, Depends, Query
+from sqlalchemy.ext.asyncio import AsyncSession

+from app.db.session import get_db
 from app.services.earth_news import get_earth_news_payload

 router = APIRouter()
@@ -9,5 +11,6 @@ router = APIRouter()
 async def get_earth_feed(
    lat: float | None = Query(None, description="Current Earth view center latitude"),
    lon: float | None = Query(None, description="Current Earth view center longitude"),
+    db: AsyncSession = Depends(get_db),
 ):
-    return await get_earth_news_payload(lat=lat, lon=lon)
+    return await get_earth_news_payload(lat=lat, lon=lon, db=db)
--- a/backend/app/api/v1/realtime_sources.py
+++ b/backend/app/api/v1/realtime_sources.py
@@ -0,0 +1,280 @@
+"""Realtime datasource operations and runtime statistics."""
+
+from datetime import UTC, datetime, timedelta
+import os
+from typing import Any
+
+from fastapi import APIRouter, Depends, HTTPException
+from sqlalchemy import distinct, func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.core.data_sources import get_data_sources_config
+from app.core.security import get_current_user
+from app.core.time import to_iso8601_utc
+from app.db.session import get_db
+from app.models.datasource import DataSource
+from app.models.datasource_config import DataSourceConfig
+from app.models.user import User
+from app.models.vessel import AISRawObservation, AISSourceHealth
+from app.services.custom_datasource_runtime import (
+    get_custom_stream_status,
+    start_custom_stream,
+    stop_custom_stream,
+)
+from app.services.scheduler import (
+    cancel_running_collector_now,
+    is_collector_running,
+    run_collector_now,
+)
+from app.services.vessel_ais_aggregation import VESSEL_AIS_SCHEMA, update_ais_source_health
+
+router = APIRouter()
+
+BUILTIN_REALTIME_SOURCES = {"aisstream_vessels"}
+REALTIME_SOURCE_TYPES = {"websocket", "ws"}
+
+
+def _is_realtime_config(config: DataSourceConfig) -> bool:
+    return str(config.source_type or "").lower() in REALTIME_SOURCE_TYPES
+
+
+def _safe_config_dict(value: Any) -> dict[str, Any]:
+    return value if isinstance(value, dict) else {}
+
+
+def _credential_configured(config: DataSourceConfig | None) -> bool:
+    if config is not None:
+        auth_config = _safe_config_dict(config.auth_config)
+        config_payload = _safe_config_dict(config.config)
+        if auth_config.get("api_key") or config_payload.get("api_key"):
+            return True
+    return bool(os.getenv("AISSTREAM_API_KEY"))
+
+
+async def _load_realtime_stats(db: AsyncSession, source: str) -> dict[str, Any]:
+    now = datetime.now(UTC)
+    observed_24h = now - timedelta(hours=24)
+    observed_1h = now - timedelta(hours=1)
+    payload_mmsi = AISRawObservation.entity_key
+
+    result = await db.execute(
+        select(
+            func.count(AISRawObservation.id).label("total_observations"),
+            func.count(AISRawObservation.id)
+            .filter(AISRawObservation.observed_at >= observed_24h)
+            .label("observations_24h"),
+            func.count(AISRawObservation.id)
+            .filter(AISRawObservation.observed_at >= observed_1h)
+            .label("observations_1h"),
+            func.count(distinct(payload_mmsi)).label("unique_mmsi_total"),
+            func.count(distinct(payload_mmsi))
+            .filter(AISRawObservation.observed_at >= observed_24h)
+            .label("unique_mmsi_24h"),
+            func.max(AISRawObservation.observed_at).label("latest_observed_at"),
+            func.max(AISRawObservation.collected_at).label("latest_collected_at"),
+        )
+        .where(AISRawObservation.target_schema == VESSEL_AIS_SCHEMA)
+        .where(AISRawObservation.source == source)
+    )
+    row = result.mappings().one()
+    return {
+        "total_observations": int(row["total_observations"] or 0),
+        "observations_24h": int(row["observations_24h"] or 0),
+        "observations_1h": int(row["observations_1h"] or 0),
+        "unique_mmsi_total": int(row["unique_mmsi_total"] or 0),
+        "unique_mmsi_24h": int(row["unique_mmsi_24h"] or 0),
+        "latest_observed_at": to_iso8601_utc(row["latest_observed_at"]),
+        "latest_collected_at": to_iso8601_utc(row["latest_collected_at"]),
+    }
+
+
+def _runtime_status_for_builtin(source: str) -> dict[str, Any]:
+    running = is_collector_running(source)
+    return {
+        "running": running,
+        "done": False,
+        "runtime": "collector",
+    }
+
+
+def _runtime_status_for_custom(config_id: int) -> dict[str, Any]:
+    status = get_custom_stream_status(config_id)
+    return {
+        "running": bool(status.get("running")),
+        "done": bool(status.get("done")),
+        "runtime": "custom_stream",
+    }
+
+
+async def _serialize_builtin_aisstream(
+    db: AsyncSession,
+    datasource: DataSource,
+    config: DataSourceConfig | None,
+) -> dict[str, Any]:
+    health = await db.get(AISSourceHealth, datasource.source)
+    config_payload = _safe_config_dict(config.config if config else {})
+    endpoint = (
+        (config.endpoint if config else None)
+        or get_data_sources_config().get_yaml_url(datasource.source)
+    )
+    return {
+        "source": datasource.source,
+        "name": datasource.name,
+        "display_name": "AISStream 实时船舶",
+        "kind": "builtin",
+        "source_type": "websocket",
+        "endpoint": endpoint,
+        "is_active": bool(datasource.is_active),
+        "credential_configured": _credential_configured(config),
+        "message_types": config_payload.get("message_types") or ["PositionReport", "ShipStaticData"],
+        "bounding_boxes": config_payload.get("bounding_boxes") or [[[-90, -180], [90, 180]]],
+        "config": config_payload,
+        "runtime": _runtime_status_for_builtin(datasource.source),
+        "health": health.to_dict() if health else None,
+        "stats": await _load_realtime_stats(db, datasource.source),
+    }
+
+
+async def _serialize_custom_stream(
+    db: AsyncSession,
+    config: DataSourceConfig,
+) -> dict[str, Any]:
+    health = await db.get(AISSourceHealth, config.name)
+    config_payload = _safe_config_dict(config.config)
+    return {
+        "source": config.name,
+        "name": config.name,
+        "display_name": config.description or config.name,
+        "kind": "custom",
+        "config_id": config.id,
+        "source_type": config.source_type,
+        "endpoint": config.endpoint,
+        "is_active": bool(config.is_active),
+        "credential_configured": config.auth_type == "none" or bool(_safe_config_dict(config.auth_config)),
+        "message_types": config_payload.get("message_types") or [],
+        "bounding_boxes": config_payload.get("bounding_boxes") or [],
+        "config": config_payload,
+        "runtime": _runtime_status_for_custom(config.id),
+        "health": health.to_dict() if health else None,
+        "stats": await _load_realtime_stats(db, config.name),
+    }
+
+
+async def _load_builtin_aisstream(db: AsyncSession) -> tuple[DataSource | None, DataSourceConfig | None]:
+    result = await db.execute(select(DataSource).where(DataSource.source == "aisstream_vessels"))
+    datasource = result.scalar_one_or_none()
+    config_result = await db.execute(
+        select(DataSourceConfig)
+        .where(DataSourceConfig.name == "aisstream_vessels")
+        .where(DataSourceConfig.is_active.is_(True))
+        .order_by(DataSourceConfig.id.desc())
+        .limit(1)
+    )
+    return datasource, config_result.scalar_one_or_none()
+
+
+async def _load_custom_realtime_config(db: AsyncSession, source: str) -> DataSourceConfig | None:
+    result = await db.execute(
+        select(DataSourceConfig)
+        .where(DataSourceConfig.name == source)
+        .order_by(DataSourceConfig.id.desc())
+        .limit(1)
+    )
+    config = result.scalar_one_or_none()
+    return config if config is not None and _is_realtime_config(config) else None
+
+
+@router.get("")
+async def list_realtime_sources(
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    sources: list[dict[str, Any]] = []
+    datasource, builtin_config = await _load_builtin_aisstream(db)
+    if datasource is not None:
+        sources.append(await _serialize_builtin_aisstream(db, datasource, builtin_config))
+
+    custom_result = await db.execute(
+        select(DataSourceConfig)
+        .where(func.lower(DataSourceConfig.source_type).in_(REALTIME_SOURCE_TYPES))
+        .order_by(DataSourceConfig.name)
+    )
+    for config in custom_result.scalars().all():
+        if config.name in BUILTIN_REALTIME_SOURCES:
+            continue
+        sources.append(await _serialize_custom_stream(db, config))
+
+    return {"total": len(sources), "data": sources}
+
+
+async def _ensure_builtin_startable(db: AsyncSession) -> DataSourceConfig | None:
+    datasource, config = await _load_builtin_aisstream(db)
+    if datasource is None:
+        raise HTTPException(status_code=404, detail="Realtime source not found")
+    if not datasource.is_active:
+        raise HTTPException(status_code=400, detail="Realtime source is disabled")
+    if not _credential_configured(config):
+        raise HTTPException(status_code=400, detail="AISStream API key is not configured")
+    return config
+
+
+@router.post("/{source}/start")
+async def start_realtime_source(
+    source: str,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    if source == "aisstream_vessels":
+        await _ensure_builtin_startable(db)
+        if is_collector_running(source):
+            return {"status": "already_running", "source": source, "runtime": _runtime_status_for_builtin(source)}
+        if not run_collector_now(source):
+            raise HTTPException(status_code=409, detail="Realtime source could not be started")
+        return {"status": "started", "source": source, "runtime": _runtime_status_for_builtin(source)}
+
+    config = await _load_custom_realtime_config(db, source)
+    if config is None:
+        raise HTTPException(status_code=404, detail="Realtime source not found")
+    if not config.is_active:
+        raise HTTPException(status_code=400, detail="Realtime source is disabled")
+    started = start_custom_stream(config.id)
+    return {
+        "status": "started" if started else "already_running",
+        "source": source,
+        "runtime": _runtime_status_for_custom(config.id),
+    }
+
+
+@router.post("/{source}/stop")
+async def stop_realtime_source(
+    source: str,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    if source == "aisstream_vessels":
+        stopped = await cancel_running_collector_now(source)
+        await update_ais_source_health(db, source=source, connection_state="disconnected", last_error=None)
+        await db.commit()
+        return {"status": "stopped" if stopped else "not_running", "source": source, "runtime": _runtime_status_for_builtin(source)}
+
+    config = await _load_custom_realtime_config(db, source)
+    if config is None:
+        raise HTTPException(status_code=404, detail="Realtime source not found")
+    stopped = await stop_custom_stream(config.id)
+    await update_ais_source_health(db, source=source, connection_state="disconnected", last_error=None)
+    await db.commit()
+    return {
+        "status": "stopped" if stopped else "not_running",
+        "source": source,
+        "runtime": _runtime_status_for_custom(config.id),
+    }
+
+
+@router.post("/{source}/restart")
+async def restart_realtime_source(
+    source: str,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    await stop_realtime_source(source, current_user=current_user, db=db)
+    return await start_realtime_source(source, current_user=current_user, db=db)
--- a/backend/app/api/v1/settings.py
+++ b/backend/app/api/v1/settings.py
@@ -15,6 +15,13 @@ from app.core.time import to_iso8601_utc
 from app.core.config import settings as app_settings
 from app.core.data_sources import get_data_sources_config
 from app.core.datasource_defaults import DEFAULT_DATASOURCES
+from app.ai_tasks.prompts import (
+    get_effective_prompt,
+    list_effective_prompts,
+    reset_prompt_override,
+    save_prompt_override,
+    serialize_effective_prompt,
+)
 from app.db.session import get_db
 from app.models.datasource import DataSource
 from app.models.datasource_config import DataSourceConfig
@@ -59,6 +66,8 @@ from app.services.scheduler import sync_datasource_job
 from app.services.tv_streams import DEFAULT_TV_SETTINGS, get_tv_settings_payload, normalize_tv_settings

 router = APIRouter()
+AI_PROVIDER_QUICK_CONNECT_TIMEOUT_SECONDS = 5
+AI_CONNECTION_TEST_PROMPT_KEY = "ai.connection_test"

 DEFAULT_SETTINGS = {
    "system": {
@@ -250,6 +259,11 @@ class OCRIntegrationUpdate(BaseModel):
    output_format: str = Field(default="markdown", pattern="^(markdown|text|json)$")


+class AIPromptUpdate(BaseModel):
+    system_prompt: str = Field(default="", max_length=8000)
+    prompt: str = Field(min_length=1, max_length=20000)
+
+
 class ExternalIntegrationsUpdate(BaseModel):
    ai_provider: AIProviderIntegrationUpdate
    barentswatch: BarentsWatchIntegrationUpdate
@@ -484,7 +498,13 @@ def _is_secret_placeholder(value: Optional[str], current_preview: str = "") -> b
    text = str(value).strip()
    if not text:
        return True
-    return text == current_preview or text.startswith("••••") or "*" in text
+    if text == current_preview or text.startswith("••••"):
+        return True
+    if "-" in text:
+        _prefix, masked = text.split("-", 1)
+        if masked and all(char in {"*", "•", " ", "\t"} for char in masked):
+            return True
+    return all(char in {"*", "•", " ", "\t"} for char in text)


 def _build_ai_provider_payload(current_payload: dict, update: AIProviderIntegrationUpdate) -> dict:
@@ -561,6 +581,56 @@ def _runtime_config_from_ai_payload(ai_payload: dict) -> dict:
    }


+def _ai_provider_runtime_fingerprint(ai_payload: dict) -> dict:
+    runtime_config = _runtime_config_from_ai_payload(ai_payload)
+    llm_config = runtime_config.get("llm_config") or {}
+    return {
+        "service_url": runtime_config.get("service_url") or "",
+        "service_token": runtime_config.get("service_token") or "",
+        "timeout_seconds": int(runtime_config.get("timeout_seconds") or 0),
+        "retry_attempts": int(runtime_config.get("retry_attempts") or 0),
+        "provider": llm_config.get("provider") or "",
+        "provider_api": llm_config.get("provider_api") or "",
+        "base_url": llm_config.get("base_url") or "",
+        "model": llm_config.get("model") or "",
+        "api_key": llm_config.get("api_key") or "",
+        "max_tokens": int(llm_config.get("max_tokens") or 0),
+        "anthropic_version": llm_config.get("anthropic_version") or "",
+    }
+
+
+async def _validate_ai_provider_full_connection(ai_payload: dict) -> dict:
+    runtime_config = _runtime_config_from_ai_payload(ai_payload)
+    client = AIProviderClient(
+        service_url=runtime_config["service_url"],
+        service_token=runtime_config["service_token"],
+        timeout=runtime_config["timeout_seconds"],
+        retry_attempts=runtime_config["retry_attempts"],
+        llm_config=runtime_config.get("llm_config") or {},
+    )
+    status_result = await client.get_status()
+    if not status_result.configured:
+        raise HTTPException(
+            status_code=400,
+            detail="AI Provider 可访问，但当前 provider/model/key 未完整配置。",
+        )
+    prompt = await get_effective_prompt(None, AI_CONNECTION_TEST_PROMPT_KEY)
+    analysis_result = await client.analyze(
+        SituationalAnalysisRequest(
+            title="保存前完整连接测试",
+            objective=prompt.prompt,
+            system_prompt=prompt.system_prompt or None,
+            observations=["这是保存 AI Provider 配置前的完整 LLM 调用测试。"],
+            constraints=["回复尽量简短。"],
+        )
+    )
+    return {
+        "status": status_result.model_dump(),
+        "provider": analysis_result.provider,
+        "model": analysis_result.model,
+    }
+
+
 def _web_search_provider_defaults(provider: str) -> dict:
    return web_search_provider_defaults(provider).model_dump()

@@ -910,7 +980,10 @@ async def save_external_integrations_payload(
    update: ExternalIntegrationsUpdate,
 ) -> dict:
    current_payload = await get_setting_payload(db, "external_integrations")
+    current_ai_payload = _normalize_ai_provider_payload(current_payload.get("ai_provider") or {})
    ai_payload = _build_ai_provider_payload(current_payload, update.ai_provider)
+    if _ai_provider_runtime_fingerprint(ai_payload) != _ai_provider_runtime_fingerprint(current_ai_payload):
+        await _validate_ai_provider_full_connection(ai_payload)
    web_search_payload = _build_web_search_payload(current_payload, update.web_search)
    ocr_payload = _build_ocr_payload(current_payload, update.ocr)

@@ -1158,6 +1231,47 @@ async def get_external_integrations(
    return {"integrations": await serialize_external_integrations(db)}


+@router.get("/ai-prompts")
+async def get_ai_prompts(
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    prompts = await list_effective_prompts(db)
+    return {"data": [serialize_effective_prompt(prompt) for prompt in prompts]}
+
+
+@router.put("/ai-prompts/{task_key}")
+async def update_ai_prompt(
+    task_key: str,
+    payload: AIPromptUpdate,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    try:
+        prompt = await save_prompt_override(
+            db,
+            task_key,
+            system_prompt=payload.system_prompt,
+            prompt=payload.prompt,
+        )
+    except KeyError:
+        raise HTTPException(status_code=404, detail="未知 AI 提示词任务") from None
+    return {"data": serialize_effective_prompt(prompt)}
+
+
+@router.post("/ai-prompts/{task_key}/reset")
+async def reset_ai_prompt(
+    task_key: str,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    try:
+        prompt = await reset_prompt_override(db, task_key)
+    except KeyError:
+        raise HTTPException(status_code=404, detail="未知 AI 提示词任务") from None
+    return {"data": serialize_effective_prompt(prompt)}
+
+
@router.get("/integrations/barentswatch/connectivity")
 async def get_barentswatch_connectivity(
    current_user: User = Depends(get_current_user),
@@ -1219,12 +1333,16 @@ async def connect_ai_provider_integration(
    current_payload = await get_setting_payload(db, "external_integrations")
    draft_ai_payload = _build_ai_provider_payload(current_payload, payload)
    runtime_config = _runtime_config_from_ai_payload(draft_ai_payload)
+    quick_llm_config = {
+        **(runtime_config.get("llm_config") or {}),
+        "max_tokens": 1,
+    }
    client = AIProviderClient(
        service_url=runtime_config["service_url"],
        service_token=runtime_config["service_token"],
-        timeout=runtime_config["timeout_seconds"],
-        retry_attempts=runtime_config["retry_attempts"],
-        llm_config=runtime_config.get("llm_config") or {},
+        timeout=min(int(runtime_config["timeout_seconds"] or 60), AI_PROVIDER_QUICK_CONNECT_TIMEOUT_SECONDS),
+        retry_attempts=1,
+        llm_config=quick_llm_config,
    )

    try:
@@ -1236,29 +1354,24 @@ async def connect_ai_provider_integration(
                "message": "AI Provider 可访问，但当前 provider/model/key 未完整配置。",
                "status": status_result.model_dump(),
            }
-        analysis_result = await client.analyze(
+        prompt = await get_effective_prompt(db, AI_CONNECTION_TEST_PROMPT_KEY)
+        probe_result = await client.analyze(
            SituationalAnalysisRequest(
-                title="连接测试",
-                objective="请用一句话回复连接可用。",
-                observations=["这是配置中心发起的 LLM 连接测试。"],
-                constraints=["回复尽量简短。"],
+                title="快速连接测试",
+                objective=prompt.prompt,
+                system_prompt=prompt.system_prompt or None,
+                observations=[],
+                constraints=["Output only OK."],
            )
        )
-        current_web_search = _normalize_web_search_payload(current_payload.get("web_search") or {})
-        current_ocr = _normalize_ocr_payload(current_payload.get("ocr") or {})
-        await save_setting_payload(
-            db,
-            "external_integrations",
-            {"ai_provider": draft_ai_payload, "web_search": current_web_search, "ocr": current_ocr},
-        )
        return {
            "success": True,
            "connected": True,
-            "message": "AI Provider 连接成功，已保存为全局默认配置。",
+            "message": "连接测试通过",
            "status": status_result.model_dump(),
-            "provider": analysis_result.provider,
-            "model": analysis_result.model,
-            "integrations": await serialize_external_integrations(db),
+            "provider": probe_result.provider,
+            "model": probe_result.model,
+            "mode": "quick_probe",
        }
    except HTTPException as exc:
        return {
--- a/backend/app/api/v1/visualization.py
+++ b/backend/app/api/v1/visualization.py
@@ -68,6 +68,7 @@ TERRAIN_TILE_BATCH_MAX_ITEMS = 128
 TERRAIN_TILE_BATCH_CONCURRENCY = 16
 _terrain_tile_cache: OrderedDict[tuple[int, int, int], tuple[bytes, str, dict[str, str]]] = OrderedDict()
 VESSEL_NAME_FALLBACK_PATTERN = re.compile(r"^mmsi\s*\d+$", re.IGNORECASE)
+VESSEL_SNAPSHOT_LEGACY_FALLBACK_ENABLED = True


 class TerrariumTileRequest(BaseModel):
@@ -941,6 +942,39 @@ def _merge_vessel_features(
    }


+async def _load_legacy_vessel_snapshot_features(
+    db: AsyncSession,
+    *,
+    bbox: tuple[float, float, float, float] | None,
+    limit: int,
+) -> list[dict[str, Any]]:
+    latest_positions = select(
+        VesselPosition.mmsi.label("mmsi"),
+        func.max(VesselPosition.received_at).label("received_at"),
+    )
+    if bbox is not None:
+        lon_min, lat_min, lon_max, lat_max = bbox
+        latest_positions = latest_positions.where(VesselPosition.lon >= lon_min)
+        latest_positions = latest_positions.where(VesselPosition.lon <= lon_max)
+        latest_positions = latest_positions.where(VesselPosition.lat >= lat_min)
+        latest_positions = latest_positions.where(VesselPosition.lat <= lat_max)
+
+    latest_positions = latest_positions.group_by(VesselPosition.mmsi).subquery()
+    result = await db.execute(
+        select(VesselPosition, VesselStatic)
+        .join(
+            latest_positions,
+            (VesselPosition.mmsi == latest_positions.c.mmsi)
+            & (VesselPosition.received_at == latest_positions.c.received_at),
+        )
+        .outerjoin(VesselStatic, VesselStatic.mmsi == VesselPosition.mmsi)
+        .order_by(VesselPosition.received_at.desc())
+        .limit(limit)
+    )
+    legacy_geojson = convert_vessels_to_geojson(list(result.all()))
+    return legacy_geojson.get("features", [])[:limit]
+
+
 def _build_vessel_stats(features: List[dict[str, Any]]) -> dict[str, Any]:
    by_type: dict[str, int] = {}
    underway = 0
@@ -1948,6 +1982,7 @@ async def collect_compute_center_location(
                provider_client=provider_client,
                query=query,
                entity_type="compute_center",
+                db=db,
                attempted_queries=attempted_queries,
                search_evidence=search_result.evidence,
            )
@@ -2073,33 +2108,6 @@ async def _load_compute_center_record(db: AsyncSession, source_id: str) -> Colle
    return result.scalars().first()


-@router.get("/geo/vessels")
-async def get_vessels_geojson(
-    bbox: Optional[str] = Query(
-        None,
-        description="Viewport bbox as lon_min,lat_min,lon_max,lat_max",
-    ),
-    type: Optional[str] = Query(
-        None,
-        description="Comma-separated vessel types: cargo,tanker,passenger,fishing,military,other",
-    ),
-    limit: Optional[int] = Query(
-        None,
-        ge=0,
-        description="Maximum vessel features to return. Omit or pass 0 for no limit.",
-    ),
-    db: AsyncSession = Depends(get_db),
-):
-    """Legacy vessel endpoint removed in favor of /api/v1/vessels/snapshot."""
-    raise HTTPException(
-        status_code=410,
-        detail=(
-            "Legacy vessel GeoJSON endpoint has been removed. "
-            "Use /api/v1/vessels/snapshot with bbox, zoom, and limit."
-        ),
-    )
-
-
 async def _load_raw_vessel_snapshot_features(
    db: AsyncSession,
    *,
@@ -2121,18 +2129,38 @@ async def _load_raw_vessel_snapshot_features(
            observed_since=observed_since,
        )
    raw_geojson = convert_aggregated_vessels_to_geojson(aggregated_vessels)
-    features = raw_geojson.get("features", [])
+    raw_features = raw_geojson.get("features", [])
+    features = raw_features
+    legacy_features: list[dict[str, Any]] = []
+    legacy_fallback_used = False
+    if not raw_features and VESSEL_SNAPSHOT_LEGACY_FALLBACK_ENABLED:
+        legacy_features = await _load_legacy_vessel_snapshot_features(
+            db,
+            bbox=bbox,
+            limit=limit,
+        )
+        features, _merge_diagnostics = _merge_vessel_features(raw_features, legacy_features)
+        legacy_fallback_used = bool(legacy_features)
+
    return features, {
-        "raw_feature_count": len(features),
+        "raw_feature_count": len(raw_features),
        "raw_unique_mmsi": len(
            {
                key
-                for key in (_feature_mmsi_key(feature) for feature in features)
+                for key in (_feature_mmsi_key(feature) for feature in raw_features)
                if key is not None
            }
        ),
-        "legacy_feature_count": 0,
-        "legacy_backfilled_mmsi": 0,
+        "legacy_feature_count": len(legacy_features),
+        "legacy_backfilled_mmsi": len(
+            {
+                key
+                for key in (_feature_mmsi_key(feature) for feature in legacy_features)
+                if key is not None
+            }
+        ),
+        "legacy_fallback_enabled": VESSEL_SNAPSHOT_LEGACY_FALLBACK_ENABLED,
+        "legacy_fallback_used": legacy_fallback_used,
        "final_unique_mmsi": len(
            {
                key
@@ -2466,7 +2494,12 @@ async def get_visualization_geo_summary(db: AsyncSession = Depends(get_db)):
        select(func.count(func.distinct(VesselPosition.mmsi)))
    )
    legacy_unique_mmsi = int(legacy_unique_result.scalar() or 0)
-    vessel_count = max(raw_unique_mmsi, legacy_unique_mmsi)
+    legacy_fallback_active = (
+        VESSEL_SNAPSHOT_LEGACY_FALLBACK_ENABLED
+        and raw_unique_mmsi == 0
+        and legacy_unique_mmsi > 0
+    )
+    vessel_count = legacy_unique_mmsi if legacy_fallback_active else raw_unique_mmsi
    aisstream_health = await db.get(AISSourceHealth, "aisstream_vessels")

    return {
@@ -2477,6 +2510,8 @@ async def get_visualization_geo_summary(db: AsyncSession = Depends(get_db)):
            "satellite_count": satellite_count,
            "compute_center_count": compute_center_count,
            "vessel_count": vessel_count,
+            "vessel_count_source": "legacy_fallback" if legacy_fallback_active else "raw_recent",
+            "vessel_legacy_fallback_enabled": VESSEL_SNAPSHOT_LEGACY_FALLBACK_ENABLED,
            "vessel_raw_unique_mmsi": raw_unique_mmsi,
            "vessel_raw_unique_window_hours": raw_unique_window_hours,
            "vessel_legacy_unique_mmsi": legacy_unique_mmsi,
--- a/backend/app/api/v1/websocket.py
+++ b/backend/app/api/v1/websocket.py
@@ -58,7 +58,7 @@ async def websocket_endpoint(

    is_anonymous = payload is None
    user_id = str(payload.get("sub")) if payload else f"anonymous:{id(websocket)}"
-    supported_channels = ["vessels"] if is_anonymous else [
+    supported_channels = ["vessels", "earth_news"] if is_anonymous else [
        "gpu_clusters",
        "submarine_cables",
        "ixp_nodes",
@@ -66,6 +66,7 @@ async def websocket_endpoint(
        "dashboard",
        "datasource_tasks",
        "vessels",
+        "earth_news",
    ]
    await manager.connect(websocket, user_id)

--- a/backend/app/core/datasource_defaults.py
+++ b/backend/app/core/datasource_defaults.py
@@ -257,6 +257,16 @@ DEFAULT_DATASOURCES = {
        "credential_provider": "aisstream",
        "credential_status": "supported",
    },
+    "media_news_archive": {
+        "id": 33,
+        "name": "Media News Archive",
+        "display_name": "媒体新闻归档",
+        "module": "L4",
+        "priority": "P2",
+        "frequency_minutes": 720,
+        "is_free": True,
+        "requires_credentials": False,
+    },
 }

 ID_TO_COLLECTOR = {info["id"]: name for name, info in DEFAULT_DATASOURCES.items()}
--- a/backend/app/db/session.py
+++ b/backend/app/db/session.py
@@ -1,6 +1,6 @@
 from typing import AsyncGenerator

-from sqlalchemy import text
+from sqlalchemy import bindparam, text
 from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine, async_sessionmaker
 from sqlalchemy.orm import declarative_base

@@ -72,6 +72,74 @@ async def seed_default_datasources(session: AsyncSession):
    await session.commit()


+LEGACY_EARTH_BOUNDARY_SOURCES = (
+    "earth_admin0_boundaries",
+    "earth_coastline",
+    "earth_claim_lines",
+    "earth_boundary_tiles",
+)
+LEGACY_EARTH_BOUNDARY_DATATYPES = (
+    "earth_boundary_source",
+    "earth_boundary_tiles",
+)
+LEGACY_EARTH_BOUNDARY_IDS = (29, 30, 31, 32)
+
+
+async def purge_legacy_earth_boundary_datasources(session: AsyncSession) -> None:
+    source_names = tuple(LEGACY_EARTH_BOUNDARY_SOURCES)
+    source_ids = tuple(LEGACY_EARTH_BOUNDARY_IDS)
+    data_types = tuple(LEGACY_EARTH_BOUNDARY_DATATYPES)
+    await session.execute(
+        text(
+            """
+            DELETE FROM datasource_mapping_templates
+            WHERE target_schema IN :data_types
+               OR datasource_config_id IN (
+                    SELECT id FROM datasource_configs WHERE name IN :source_names
+               )
+            """
+        ).bindparams(bindparam("source_names", expanding=True), bindparam("data_types", expanding=True)),
+        {"source_names": list(source_names), "data_types": list(data_types)},
+    )
+    await session.execute(
+        text("DELETE FROM datasource_configs WHERE name IN :source_names").bindparams(
+            bindparam("source_names", expanding=True)
+        ),
+        {"source_names": list(source_names)},
+    )
+    await session.execute(
+        text(
+            """
+            DELETE FROM collected_data
+            WHERE source IN :source_names OR data_type IN :data_types
+            """
+        ).bindparams(bindparam("source_names", expanding=True), bindparam("data_types", expanding=True)),
+        {"source_names": list(source_names), "data_types": list(data_types)},
+    )
+    await session.execute(
+        text(
+            """
+            DELETE FROM data_snapshots
+            WHERE source IN :source_names OR datasource_id IN :source_ids
+            """
+        ).bindparams(bindparam("source_names", expanding=True), bindparam("source_ids", expanding=True)),
+        {"source_names": list(source_names), "source_ids": list(source_ids)},
+    )
+    await session.execute(
+        text("DELETE FROM collection_tasks WHERE datasource_id IN :source_ids").bindparams(
+            bindparam("source_ids", expanding=True)
+        ),
+        {"source_ids": list(source_ids)},
+    )
+    await session.execute(
+        text("DELETE FROM data_sources WHERE source IN :source_names OR id IN :source_ids").bindparams(
+            bindparam("source_names", expanding=True), bindparam("source_ids", expanding=True)
+        ),
+        {"source_names": list(source_names), "source_ids": list(source_ids)},
+    )
+    await session.commit()
+
+
 DEFAULT_LOGIN_USERS = (
    {
        "username": "admin",
@@ -134,6 +202,7 @@ async def init_db():
    import app.models.vessel  # noqa: F401
    import app.models.vessel_enrichment  # noqa: F401
    import app.models.datasource_mapping  # noqa: F401
+    import app.models.earth_news  # noqa: F401

    logger.warning_event(
        "Database pool settings active",
@@ -202,6 +271,18 @@ async def init_db():
                """
            )
        )
+        await conn.execute(
+            text(
+                """
+                ALTER TABLE earth_news_items
+                ADD COLUMN IF NOT EXISTS content_language VARCHAR(32) NOT NULL DEFAULT 'en',
+                ADD COLUMN IF NOT EXISTS localizations JSONB NOT NULL DEFAULT '{}'::jsonb,
+                ADD COLUMN IF NOT EXISTS enrichment_status VARCHAR(80) NOT NULL DEFAULT 'pending',
+                ADD COLUMN IF NOT EXISTS enrichment_error TEXT,
+                ADD COLUMN IF NOT EXISTS enriched_at TIMESTAMPTZ
+                """
+            )
+        )
        await conn.execute(
            text(
                """
@@ -210,6 +291,22 @@ async def init_db():
                """
            )
        )
+        await conn.execute(
+            text(
+                """
+                CREATE INDEX IF NOT EXISTS idx_earth_news_enrichment_status
+                ON earth_news_items (enrichment_status)
+                """
+            )
+        )
+        await conn.execute(
+            text(
+                """
+                CREATE INDEX IF NOT EXISTS idx_earth_news_enriched_at
+                ON earth_news_items (enriched_at)
+                """
+            )
+        )
        await conn.execute(
            text(
                """
@@ -284,4 +381,5 @@ async def init_db():
        await seed_default_bgp_collector_locations(session)
        await seed_compute_center_locations_from_source_coords(session)
        await seed_default_datasources(session)
+        await purge_legacy_earth_boundary_datasources(session)
        await ensure_default_admin_user(session)
--- a/backend/app/main.py
+++ b/backend/app/main.py
@@ -1,8 +1,10 @@
 from contextlib import asynccontextmanager
+from pathlib import Path
 from uuid import uuid4

 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
 from starlette.middleware.base import BaseHTTPMiddleware

 from app.api.main import api_router
@@ -18,6 +20,10 @@ from app.services.scheduler import (
    stop_scheduler,
    sync_scheduler_with_datasources,
 )
+from app.services.earth_news_worker import (
+    start_earth_news_target_worker,
+    stop_earth_news_target_worker,
+)


 configure_logging()
@@ -53,7 +59,9 @@ async def lifespan(app: FastAPI):
    start_scheduler()
    await sync_scheduler_with_datasources()
    broadcaster.start()
+    start_earth_news_target_worker()
    yield
+    await stop_earth_news_target_worker()
    broadcaster.stop()
    stop_scheduler()

@@ -82,6 +90,14 @@ app.add_middleware(WebSocketCORSMiddleware)
 app.include_router(api_router, prefix="/api/v1")
 app.include_router(websocket.router)

+EARTH_BRAND_ASSET_DIR = Path(__file__).resolve().parents[2] / "data" / "earth-brand"
+EARTH_BRAND_ASSET_DIR.mkdir(parents=True, exist_ok=True)
+app.mount(
+    "/earth-brand-assets",
+    StaticFiles(directory=str(EARTH_BRAND_ASSET_DIR)),
+    name="earth-brand-assets",
+)
+

@app.get("/health")
 async def health_check():
--- a/backend/app/models/init.py
+++ b/backend/app/models/init.py
@@ -16,6 +16,7 @@ from app.models.playground_message import PlaygroundMessage
 from app.models.system_log import SystemLog, AuditLog
 from app.models.vessel import AISConflictRecord, AISRawObservation, AISSourceHealth, VesselPosition, VesselStatic
 from app.models.datasource_mapping import DataSourceMappingTemplate
+from app.models.earth_news import EarthNewsItem

 __all__ = [
    "User",
@@ -43,4 +44,5 @@ __all__ = [
    "AISConflictRecord",
    "AISSourceHealth",
    "DataSourceMappingTemplate",
+    "EarthNewsItem",
 ]
--- a/backend/app/models/earth_news.py
+++ b/backend/app/models/earth_news.py
@@ -0,0 +1,40 @@
+from sqlalchemy import Boolean, Column, DateTime, Float, Index, JSON, String, Text
+from sqlalchemy.sql import func
+
+from app.db.session import Base
+
+
+class EarthNewsItem(Base):
+    __tablename__ = "earth_news_items"
+
+    id = Column(String(160), primary_key=True)
+    title = Column(String(500), nullable=False)
+    summary = Column(Text, nullable=False, default="")
+    content_language = Column(String(32), nullable=False, default="en")
+    localizations = Column(JSON, nullable=False, default=dict)
+    url = Column(Text, nullable=False)
+    source = Column(String(255), nullable=False, default="")
+    feed_name = Column(String(255), nullable=False, default="")
+    region = Column(String(80), nullable=False, index=True)
+    homepage_url = Column(Text, nullable=False, default="")
+    published_at = Column(DateTime(timezone=True), nullable=True, index=True)
+
+    latitude = Column(Float, nullable=False)
+    longitude = Column(Float, nullable=False)
+    location_label = Column(String(255), nullable=False)
+    location_source = Column(String(80), nullable=False, default="region_anchor")
+    verified = Column(Boolean, nullable=False, default=False, index=True)
+    location_meta = Column(JSON, nullable=False, default=dict)
+
+    first_seen_at = Column(DateTime(timezone=True), server_default=func.now(), nullable=False)
+    last_seen_at = Column(DateTime(timezone=True), server_default=func.now(), nullable=False, index=True)
+    resolved_at = Column(DateTime(timezone=True), nullable=True, index=True)
+    enrichment_status = Column(String(80), nullable=False, default="pending", index=True)
+    enrichment_error = Column(Text, nullable=True)
+    enriched_at = Column(DateTime(timezone=True), nullable=True, index=True)
+    updated_at = Column(DateTime(timezone=True), server_default=func.now(), onupdate=func.now())
+
+    __table_args__ = (
+        Index("idx_earth_news_region_published", "region", "published_at"),
+        Index("idx_earth_news_region_seen", "region", "last_seen_at"),
+    )
--- a/backend/app/schemas/ai.py
+++ b/backend/app/schemas/ai.py
@@ -13,10 +13,11 @@ class AIContentBlock(BaseModel):

 class SituationalAnalysisRequest(BaseModel):
    title: str = Field(..., min_length=1, max_length=200)
-    objective: str = Field(..., min_length=1, max_length=1000)
+    objective: str = Field(..., min_length=1, max_length=20000)
    context: dict[str, Any] = Field(default_factory=dict)
    observations: list[str] = Field(default_factory=list)
    constraints: list[str] = Field(default_factory=list)
+    system_prompt: str | None = Field(default=None, max_length=8000)
    preferred_model: str | None = Field(default=None, max_length=200)
    thinking: dict[str, Any] | None = None

--- a/backend/app/services/alert_ai_brief.py
+++ b/backend/app/services/alert_ai_brief.py
@@ -8,6 +8,9 @@ from sqlalchemy.ext.asyncio import AsyncSession

 from app.models.alert import Alert, AlertSeverity, AlertStatus
 from app.schemas.ai import AlertBriefRequest, SituationalAnalysisRequest
+from app.ai_tasks.prompts import get_effective_prompt
+
+ALERT_BRIEF_PROMPT_KEY = "alerts.brief"


 def _format_counter(counter: Counter[str], empty_text: str = "无") -> str:
@@ -84,11 +87,13 @@ async def build_alert_brief_request(
        "top_datasources": dict(datasource_counts.most_common(6)),
        "top_active_datasources": dict(active_datasource_counts.most_common(5)),
    }
+    prompt = await get_effective_prompt(db, ALERT_BRIEF_PROMPT_KEY)

    return (
        SituationalAnalysisRequest(
            title="告警态势 AI 简报",
-            objective="基于当前告警总量、严重度、状态、数据源分布与最近告警摘录，生成一份面向值班人员的简明告警态势简报，突出待处理风险、告警集中点和优先动作。",
+            objective=prompt.prompt,
+            system_prompt=prompt.system_prompt or None,
            observations=facts,
            constraints=[
                "明确区分事实、推断与建议。",
--- a/backend/app/services/bgp_ai_brief.py
+++ b/backend/app/services/bgp_ai_brief.py
@@ -11,9 +11,12 @@ from app.models.bgp_anomaly import BGPAnomaly
 from app.models.bgp_incident import BGPIncident
 from app.models.bgp_observation import BGPObservation
 from app.schemas.ai import SituationalAnalysisRequest
+from app.ai_tasks.prompts import get_effective_prompt
 from app.services.bgp_collectors import build_bgp_collector_coverage
 from app.services.bgp_enrichment import lookup_prefix_geography

+BGP_BRIEF_PROMPT_KEY = "bgp.brief"
+

 def _format_counter(counter: dict[str, int], empty_text: str = "无") -> str:
    if not counter:
@@ -243,12 +246,15 @@ async def build_bgp_brief_request(
            for prefix, item in list(prefix_geographies.items())[:8]
        },
    }
+    prompt = await get_effective_prompt(db, BGP_BRIEF_PROMPT_KEY)

    return SituationalAnalysisRequest(
        title="BGP 态势 AI 简报",
-        objective="基于当前 BGP incidents、anomalies、原始观测事件、观测站覆盖与 prefix geography 证据，生成一份面向操作员的简明态势简报，突出区域热点、观测偏差、当前风险、证据和优先动作。",
+        objective=prompt.prompt,
+        system_prompt=prompt.system_prompt or None,
        observations=observations_lines,
        constraints=[
+            "直接输出中文 Markdown 简报正文，不要输出英文写作计划、提示词复述、字段说明或元评论。",
            "明确区分事实、推断与建议。",
            "优先指出需要立即关注的高严重度 incident 或异常模式。",
            "需要单独指出哪些区域结论来自 prefix geography / affected regions，哪些可能受 collector coverage 偏差影响。",
--- a/backend/app/services/collectors/init.py
+++ b/backend/app/services/collectors/init.py
@@ -36,6 +36,7 @@ from app.services.collectors.iptoasn import IPtoASNPrefixGeoCollector
 from app.services.collectors.opengeofeed import OpenGeoFeedPrefixGeoCollector
 from app.services.collectors.nro_delegated import NRODelegatedPrefixGeoCollector
 from app.services.collectors.news_live_streams import NewsLiveStreamsCollector
+from app.services.collectors.media_news_archive import MediaNewsArchiveCollector
 from app.services.collectors.aisstream import AISStreamCollector
 from app.services.collectors.vessel_ais import VesselAISCollector

@@ -65,6 +66,7 @@ collector_registry.register(IPtoASNPrefixGeoCollector())
 collector_registry.register(OpenGeoFeedPrefixGeoCollector())
 collector_registry.register(NRODelegatedPrefixGeoCollector())
 collector_registry.register(NewsLiveStreamsCollector())
+collector_registry.register(MediaNewsArchiveCollector())
 collector_registry.register(VesselAISCollector())
 collector_registry.register(AISStreamCollector())

@@ -100,6 +102,7 @@ __all__ = [
    "OpenGeoFeedPrefixGeoCollector",
    "NRODelegatedPrefixGeoCollector",
    "NewsLiveStreamsCollector",
+    "MediaNewsArchiveCollector",
    "VesselAISCollector",
    "AISStreamCollector",
 ]
--- a/backend/app/services/collectors/media_news_archive.py
+++ b/backend/app/services/collectors/media_news_archive.py
@@ -0,0 +1,57 @@
+from __future__ import annotations
+
+from typing import Any
+
+from app.services.collectors.base import BaseCollector
+from app.services.earth_news_store import list_all_earth_news_records
+
+
+class MediaNewsArchiveCollector(BaseCollector):
+    name = "media_news_archive"
+    priority = "P2"
+    module = "L4"
+    frequency_hours = 12
+    data_type = "news_item"
+    fail_on_empty = False
+
+    async def fetch(self) -> list[dict[str, Any]]:
+        if not self._db_session:
+            return []
+
+        records = await list_all_earth_news_records(self._db_session)
+        items: list[dict[str, Any]] = []
+        for record in records:
+            location_meta = dict(record.location_meta or {})
+            target = location_meta.get("target") if isinstance(location_meta.get("target"), dict) else {}
+            country = target.get("country")
+            city = target.get("city")
+            items.append(
+                {
+                    "id": record.id,
+                    "source_id": record.id,
+                    "name": record.title,
+                    "title": record.title,
+                    "description": record.summary,
+                    "country": country,
+                    "city": city,
+                    "latitude": record.latitude,
+                    "longitude": record.longitude,
+                    "reference_date": record.published_at,
+                    "metadata": {
+                        "url": record.url,
+                        "source": record.source,
+                        "feed_name": record.feed_name,
+                        "region": record.region,
+                        "homepage_url": record.homepage_url,
+                        "published_at": record.published_at.isoformat() if record.published_at else None,
+                        "location_label": record.location_label,
+                        "location_source": record.location_source,
+                        "verified": record.verified,
+                        "location_meta": location_meta,
+                        "first_seen_at": record.first_seen_at.isoformat() if record.first_seen_at else None,
+                        "last_seen_at": record.last_seen_at.isoformat() if record.last_seen_at else None,
+                        "resolved_at": record.resolved_at.isoformat() if record.resolved_at else None,
+                    },
+                }
+            )
+        return items
--- a/backend/app/services/credential_guides.py
+++ b/backend/app/services/credential_guides.py
@@ -7,6 +7,7 @@ from typing import Any

 from sqlalchemy import select

+from app.ai_tasks.prompts import get_effective_prompt
 from app.models.system_setting import SystemSetting
 from app.schemas.ai import SituationalAnalysisRequest
 from app.services.ai_client import AIProviderClient
@@ -15,6 +16,7 @@ from app.services.ai_tools.web_search import WebSearchClient, WebSearchError


 CREDENTIAL_GUIDES_CATEGORY = "collector_credential_guides"
+CREDENTIAL_GUIDE_PROMPT_KEY = "credential.guide"


@dataclass(frozen=True)
@@ -240,14 +242,12 @@ async def generate_credential_guide(
        guide["sources"] = []
        return guide

+    prompt = await get_effective_prompt(db, CREDENTIAL_GUIDE_PROMPT_KEY)
    response = await ai_client.analyze(
        SituationalAnalysisRequest(
            title=f"Generate credential guide for {provider}",
-            objective=(
-                default.prompt
-                + "\n只能根据 context.search_evidence 中的来源生成教程；"
-                + "如果证据不足，明确说明需要以官方页面为准。"
-            ),
+            objective=f"{default.prompt}\n{prompt.prompt}",
+            system_prompt=prompt.system_prompt or None,
            context={
                "provider": provider,
                "current_default_guide": default.markdown,
--- a/backend/app/services/datasource_mapping.py
+++ b/backend/app/services/datasource_mapping.py
@@ -8,6 +8,7 @@ import re
 from datetime import UTC, datetime
 from typing import Any

+from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession

 from app.core.target_schema_registry import TargetSchema, get_target_schema
@@ -254,6 +255,12 @@ def _best_field_match(field_name: str, candidates: list[str]) -> str | None:
        "lat": ("lat", "latitude", "y"),
        "lon": ("lon", "lng", "longitude", "x"),
        "mmsi": ("mmsi",),
+        "geometry": ("geometry", "geom"),
+        "properties": ("properties", "props"),
+        "source_kind": ("source_kind", "kind", "type"),
+        "feature_count": ("feature_count", "features_count", "count"),
+        "artifact_path": ("artifact_path", "path", "file"),
+        "sha256": ("sha256", "hash", "checksum"),
        "sog": ("sog", "speed", "speedOverGround"),
        "cog": ("cog", "course", "courseOverGround"),
        "received_at": ("received_at", "timestamp", "time", "updated_at"),
--- a/backend/app/services/earth_boundaries.py
+++ b/backend/app/services/earth_boundaries.py
@@ -0,0 +1,671 @@
+"""Earth boundary static asset service."""
+
+from __future__ import annotations
+
+import asyncio
+import hashlib
+import json
+import shutil
+import sys
+from datetime import UTC, datetime
+from pathlib import Path
+from typing import Any
+from uuid import uuid4
+
+import httpx
+
+
+REPO_ROOT = Path(__file__).resolve().parents[3]
+SOURCE_OUTPUT_DIR = REPO_ROOT / "data/earth-boundary-sources"
+SOURCE_MANIFEST_PATH = SOURCE_OUTPUT_DIR / "manifest.json"
+BUILD_RESULT_PATH = SOURCE_OUTPUT_DIR / "build-result.json"
+BUILD_JOB_PATH = SOURCE_OUTPUT_DIR / "build-job.json"
+BOUNDARY_OUTPUT_DIR = REPO_ROOT / "frontend/public/earth/data/boundaries/v1"
+BOUNDARY_MANIFEST_PATH = BOUNDARY_OUTPUT_DIR / "manifest.json"
+PMTILES_ARTIFACT_PATH = (
+    REPO_ROOT / "frontend/public/earth/data/boundaries/earth-boundaries-china-pov-v1.pmtiles"
+)
+LEGACY_GEOJSON_PATH = REPO_ROOT / "frontend/public/earth/data/countries-admin0.min.geojson"
+POV_POLICY_PATH = REPO_ROOT / "config/earth-boundary-pov-policy.china-v1.json"
+LOCAL_CONFIG_PATH = REPO_ROOT / "config/earth-boundary-sources.local.json"
+EXAMPLE_CONFIG_PATH = REPO_ROOT / "config/earth-boundary-sources.example.json"
+
+BOUNDARY_SOURCE_KINDS = {
+    "earth_admin0_boundaries": "admin0-boundaries",
+    "earth_coastline": "coastline",
+    "earth_claim_lines": "claim-lines",
+}
+
+DEFAULT_PUBLIC_BOUNDARY_SOURCES = {
+    "earth_admin0_boundaries": {
+        "displayName": "Natural Earth Admin-0 Countries",
+        "sourceKind": "admin0-boundaries",
+        "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_admin_0_countries.geojson",
+        "method": "GET",
+        "headers": {},
+        "auth_type": "none",
+        "license": "Natural Earth public domain",
+    },
+    "earth_coastline": {
+        "displayName": "Natural Earth Coastline",
+        "sourceKind": "coastline",
+        "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_coastline.geojson",
+        "method": "GET",
+        "headers": {},
+        "auth_type": "none",
+        "license": "Natural Earth public domain",
+    },
+    "earth_claim_lines": {
+        "displayName": "Natural Earth Disputed Boundaries",
+        "sourceKind": "claim-lines",
+        "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_admin_0_boundary_lines_disputed_areas.geojson",
+        "method": "GET",
+        "headers": {},
+        "auth_type": "none",
+        "license": "Natural Earth public domain",
+    },
+}
+
+BUILD_CONFIG = {
+    "builder": "scripts/build_earth_boundary_pmtiles.py",
+    "format": "pmtiles+mvt",
+    "production_target": "pmtiles-mvt",
+}
+
+
+class EarthBoundaryBuildError(RuntimeError):
+    def __init__(self, message: str, *, code: str = "build_failed", details: Any = None) -> None:
+        super().__init__(message)
+        self.code = code
+        self.details = details
+
+
+_build_job_lock = asyncio.Lock()
+_build_task: asyncio.Task | None = None
+_build_job_state: dict[str, Any] = {}
+
+
+def _utc_now_iso() -> str:
+    return datetime.now(UTC).isoformat()
+
+
+def _public_job_state() -> dict[str, Any]:
+    if _build_job_state:
+        return dict(_build_job_state)
+    return _read_json(BUILD_JOB_PATH)
+
+
+def get_boundary_build_status() -> dict[str, Any]:
+    return {"job": _public_job_state()}
+
+
+def _set_job_state(**updates: Any) -> dict[str, Any]:
+    global _build_job_state
+    current = dict(_build_job_state)
+    current.update(updates)
+    current["updated_at"] = _utc_now_iso()
+    _build_job_state = current
+    _write_json(BUILD_JOB_PATH, current)
+    return current
+
+
+def _append_job_log(message: str) -> None:
+    logs = list(_build_job_state.get("logs") or [])
+    logs.append({"time": _utc_now_iso(), "message": message})
+    _set_job_state(logs=logs[-40:])
+
+
+def _update_job_progress(progress: float, phase: str, message: str, **extra: Any) -> None:
+    bounded_progress = max(0, min(100, int(round(progress))))
+    _set_job_state(
+        status="running",
+        progress=bounded_progress,
+        phase=phase,
+        message=message,
+        **extra,
+    )
+
+
+def _read_json(path: Path) -> dict[str, Any]:
+    if not path.exists():
+        return {}
+    with path.open("r", encoding="utf-8") as f:
+        payload = json.load(f)
+    return payload if isinstance(payload, dict) else {}
+
+
+def _write_json(path: Path, payload: dict[str, Any]) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    with path.open("w", encoding="utf-8") as f:
+        json.dump(payload, f, ensure_ascii=False, indent=2)
+        f.write("\n")
+
+
+def _sha256_bytes(payload: bytes) -> str:
+    return hashlib.sha256(payload).hexdigest()
+
+
+def _stable_json_hash(payload: Any) -> str:
+    encoded = json.dumps(payload, ensure_ascii=False, sort_keys=True, separators=(",", ":"))
+    return hashlib.sha256(encoded.encode("utf-8")).hexdigest()
+
+
+def _artifact_extension(endpoint: str, content_type: str, payload: bytes) -> str:
+    suffix = Path(endpoint.split("?", 1)[0]).suffix.lower()
+    if suffix in {".json", ".geojson", ".zip", ".pbf"}:
+        return suffix
+    if "geo+json" in content_type or b'"FeatureCollection"' in payload[:4096]:
+        return ".geojson"
+    if "json" in content_type:
+        return ".json"
+    return ".dat"
+
+
+def _json_feature_count(payload: Any) -> int:
+    if isinstance(payload, dict) and isinstance(payload.get("features"), list):
+        return len(payload["features"])
+    if isinstance(payload, list):
+        return len(payload)
+    return 1 if payload else 0
+
+
+def _directory_stats(path: Path) -> dict[str, int]:
+    if not path.exists():
+        return {"file_count": 0, "size_bytes": 0}
+    files = [item for item in path.rglob("*") if item.is_file()]
+    return {"file_count": len(files), "size_bytes": sum(item.stat().st_size for item in files)}
+
+
+def _load_source_feature_collection(source: dict[str, Any]) -> dict[str, Any]:
+    path = REPO_ROOT / source["path"]
+    payload = _read_json(path)
+    features = payload.get("features") if isinstance(payload, dict) else None
+    return {
+        "type": "FeatureCollection",
+        "features": features if isinstance(features, list) else [],
+    }
+
+
+def _write_high_precision_geojson_manifest(
+    sources: list[dict[str, Any]],
+    build_input_hash: str,
+    missing_tools: list[str],
+) -> dict[str, Any]:
+    BOUNDARY_OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
+    admin0 = next(source for source in sources if source["kind"] == "admin0-boundaries")
+    coastline = next(source for source in sources if source["kind"] == "coastline")
+    claim_lines = next(source for source in sources if source["kind"] == "claim-lines")
+
+    admin0_payload = _load_source_feature_collection(admin0)
+    coastline_payload = _load_source_feature_collection(coastline)
+    claim_payload = _load_source_feature_collection(claim_lines)
+    for feature in coastline_payload["features"]:
+        props = feature.setdefault("properties", {})
+        if isinstance(props, dict):
+            props["PLANET_LAYER"] = "coastline"
+
+    base_payload = {
+        "type": "FeatureCollection",
+        "features": [*admin0_payload["features"], *coastline_payload["features"]],
+    }
+    base_path = BOUNDARY_OUTPUT_DIR / "earth-boundaries-high-precision-base.geojson"
+    hover_path = BOUNDARY_OUTPUT_DIR / "earth-boundaries-high-precision-hover.geojson"
+    claim_path = BOUNDARY_OUTPUT_DIR / "earth-boundaries-high-precision-claims.geojson"
+    _write_json(base_path, base_payload)
+    _write_json(hover_path, admin0_payload)
+    _write_json(claim_path, claim_payload)
+
+    manifest = {
+        "version": "natural-earth-v1",
+        "builtAt": _utc_now_iso(),
+        "tileProvider": "geojson-high-precision",
+        "format": "geojson-directory",
+        "buildInputHash": build_input_hash,
+        "base": base_path.name,
+        "hoverIndex": hover_path.name,
+        "claimLine": claim_path.name,
+        "sourceFeatureCount": {
+            "admin0": len(admin0_payload["features"]),
+            "coastline": len(coastline_payload["features"]),
+            "claimLines": len(claim_payload["features"]),
+        },
+        "pmtiles": None,
+        "missingTools": missing_tools,
+    }
+    _write_json(BOUNDARY_MANIFEST_PATH, manifest)
+    return manifest
+
+
+def _relative(path: Path) -> str:
+    return str(path.relative_to(REPO_ROOT))
+
+
+def load_boundary_config() -> tuple[dict[str, Any], str]:
+    if LOCAL_CONFIG_PATH.exists():
+        return _read_json(LOCAL_CONFIG_PATH), "local"
+    return _read_json(EXAMPLE_CONFIG_PATH), "example"
+
+
+def save_boundary_config(payload: dict[str, Any]) -> dict[str, Any]:
+    if not isinstance(payload, dict):
+        raise EarthBoundaryBuildError("Earth boundary config must be a JSON object", code="invalid_config")
+    _write_json(LOCAL_CONFIG_PATH, payload)
+    return get_boundary_status()
+
+
+def _source_configs(payload: dict[str, Any]) -> dict[str, Any]:
+    raw_sources = payload.get("collectorConfigs") or payload.get("sources") or {}
+    return raw_sources if isinstance(raw_sources, dict) else {}
+
+
+def _is_placeholder_endpoint(endpoint: Any) -> bool:
+    value = str(endpoint or "").strip()
+    return not value or "example.com" in value
+
+
+def _source_configs_with_defaults(payload: dict[str, Any]) -> dict[str, Any]:
+    raw_sources = _source_configs(payload)
+    merged: dict[str, Any] = {}
+    for source_key, default_config in DEFAULT_PUBLIC_BOUNDARY_SOURCES.items():
+        configured = raw_sources.get(source_key)
+        if not isinstance(configured, dict) or _is_placeholder_endpoint(configured.get("endpoint")):
+            merged[source_key] = dict(default_config)
+        else:
+            merged[source_key] = {**default_config, **configured}
+    for source_key, source_config in raw_sources.items():
+        if source_key not in merged:
+            merged[source_key] = source_config
+    return merged
+
+
+def _build_input_hash(source_manifest: dict[str, Any]) -> str:
+    return _stable_json_hash(
+        {
+            "source_manifest_schema": source_manifest.get("schema"),
+            "sources": [
+                {
+                    "id": source.get("id"),
+                    "sha256": source.get("sha256"),
+                    "kind": source.get("kind"),
+                }
+                for source in source_manifest.get("sources", [])
+            ],
+            "pov_policy": source_manifest.get("povPolicy"),
+            "build_config": BUILD_CONFIG,
+        }
+    )
+
+
+def _has_current_artifacts(boundary_manifest: dict[str, Any], build_input_hash: str) -> bool:
+    return (
+        bool(boundary_manifest)
+        and boundary_manifest.get("buildInputHash") == build_input_hash
+        and boundary_manifest.get("tileProvider") == "pmtiles-mvt"
+        and PMTILES_ARTIFACT_PATH.exists()
+    )
+
+
+def get_boundary_status() -> dict[str, Any]:
+    config_payload, config_source = load_boundary_config()
+    effective_source_configs = _source_configs_with_defaults(config_payload)
+    boundary_manifest = _read_json(BOUNDARY_MANIFEST_PATH)
+    pmtiles_exists = PMTILES_ARTIFACT_PATH.exists()
+    manifest_exists = BOUNDARY_MANIFEST_PATH.exists()
+    high_precision_ready = (
+        manifest_exists
+        and (
+            (
+                boundary_manifest.get("tileProvider") == "pmtiles-mvt"
+                and pmtiles_exists
+            )
+            or boundary_manifest.get("tileProvider") == "geojson-high-precision"
+        )
+    )
+    legacy_exists = LEGACY_GEOJSON_PATH.exists()
+    provider = (
+        boundary_manifest.get("tileProvider")
+        if high_precision_ready
+        else "legacy-geojson" if legacy_exists else "missing"
+    )
+    return {
+        "provider": provider,
+        "high_precision_ready": high_precision_ready,
+        "fallback_available": legacy_exists,
+        "config_source": config_source,
+        "config_path": _relative(LOCAL_CONFIG_PATH),
+        "config_exists": LOCAL_CONFIG_PATH.exists(),
+        "config": config_payload,
+        "effective_default_sources": [
+            source_key
+            for source_key, source_config in effective_source_configs.items()
+            if source_key in DEFAULT_PUBLIC_BOUNDARY_SOURCES
+            and source_config.get("endpoint") == DEFAULT_PUBLIC_BOUNDARY_SOURCES[source_key]["endpoint"]
+        ],
+        "manifest": {
+            "path": _relative(BOUNDARY_MANIFEST_PATH),
+            "exists": manifest_exists,
+            "tileProvider": boundary_manifest.get("tileProvider"),
+            "buildInputHash": boundary_manifest.get("buildInputHash"),
+            "builtAt": boundary_manifest.get("builtAt"),
+        },
+        "pmtiles": {
+            "path": _relative(PMTILES_ARTIFACT_PATH),
+            "exists": pmtiles_exists,
+            "size_bytes": PMTILES_ARTIFACT_PATH.stat().st_size if pmtiles_exists else 0,
+        },
+        "legacy": {
+            "path": _relative(LEGACY_GEOJSON_PATH),
+            "exists": legacy_exists,
+            "size_bytes": LEGACY_GEOJSON_PATH.stat().st_size if legacy_exists else 0,
+        },
+        "source_manifest": {
+            "path": _relative(SOURCE_MANIFEST_PATH),
+            "exists": SOURCE_MANIFEST_PATH.exists(),
+        },
+        "last_build": _read_json(BUILD_RESULT_PATH),
+        "current_job": _public_job_state(),
+    }
+
+
+async def _download_source(
+    source_key: str,
+    source_config: dict[str, Any],
+    progress_callback: Any = None,
+) -> dict[str, Any]:
+    endpoint = str(source_config.get("endpoint") or "").strip()
+    if _is_placeholder_endpoint(endpoint):
+        raise EarthBoundaryBuildError(
+            f"{source_key} endpoint is not configured",
+            code="source_not_configured",
+            details={"source": source_key},
+        )
+    method = str(source_config.get("method") or "GET").upper()
+    if method not in {"GET", "POST"}:
+        raise EarthBoundaryBuildError(
+            f"{source_key} method must be GET or POST",
+            code="invalid_config",
+            details={"source": source_key, "method": method},
+        )
+
+    if endpoint.startswith("file://") or Path(endpoint).expanduser().exists():
+        payload = Path(endpoint.removeprefix("file://")).expanduser().read_bytes()
+        content_type = "application/octet-stream"
+        if progress_callback:
+            progress_callback(1, len(payload), len(payload))
+    else:
+        timeout = float(source_config.get("timeout") or 120)
+        headers = source_config.get("headers") if isinstance(source_config.get("headers"), dict) else {}
+        async with httpx.AsyncClient(timeout=timeout, follow_redirects=True) as client:
+            async with client.stream(method, endpoint, headers=headers) as response:
+                response.raise_for_status()
+                content_type = response.headers.get("content-type", "")
+                total = int(response.headers.get("content-length") or 0)
+                chunks = []
+                downloaded = 0
+                async for chunk in response.aiter_bytes():
+                    if not chunk:
+                        continue
+                    chunks.append(chunk)
+                    downloaded += len(chunk)
+                    if progress_callback:
+                        progress_callback(
+                            (downloaded / total) if total else None,
+                            downloaded,
+                            total,
+                        )
+                payload = b"".join(chunks)
+
+    extension = _artifact_extension(endpoint, content_type, payload)
+    parsed: Any = None
+    if extension in {".json", ".geojson"}:
+        parsed = json.loads(payload.decode("utf-8"))
+    feature_count = _json_feature_count(parsed)
+    if feature_count <= 0:
+        raise EarthBoundaryBuildError(
+            f"{source_key} downloaded payload contains no features",
+            code="empty_source",
+            details={"source": source_key},
+        )
+
+    sha256 = _sha256_bytes(payload)
+    source_dir = SOURCE_OUTPUT_DIR / source_key
+    source_dir.mkdir(parents=True, exist_ok=True)
+    artifact_path = source_dir / f"{sha256}{extension}"
+    artifact_path.write_bytes(payload)
+    return {
+        "id": source_key,
+        "kind": source_config.get("sourceKind") or BOUNDARY_SOURCE_KINDS[source_key],
+        "path": _relative(artifact_path),
+        "sha256": sha256,
+        "featureCount": feature_count,
+        "license": source_config.get("license"),
+    }
+
+
+async def _run_step(args: list[str]) -> dict[str, Any]:
+    process = await asyncio.create_subprocess_exec(
+        sys.executable,
+        *args,
+        cwd=REPO_ROOT,
+        stdout=asyncio.subprocess.PIPE,
+        stderr=asyncio.subprocess.PIPE,
+    )
+    stdout_bytes, stderr_bytes = await process.communicate()
+    stdout = stdout_bytes.decode("utf-8", errors="replace").strip()
+    stderr = stderr_bytes.decode("utf-8", errors="replace").strip()
+    payload: dict[str, Any] = {"stdout": stdout, "stderr": stderr, "returncode": process.returncode}
+    last_line = stdout.splitlines()[-1:] or []
+    if last_line:
+        try:
+            payload["result"] = json.loads(last_line[0])
+        except json.JSONDecodeError:
+            payload["result"] = last_line[0]
+    if process.returncode != 0:
+        raise EarthBoundaryBuildError(
+            stderr or stdout or f"command failed: {' '.join(args)}",
+            code="build_command_failed",
+            details=payload,
+        )
+    return payload
+
+
+async def build_boundary_assets(progress_callback: Any = None) -> dict[str, Any]:
+    config_payload, config_source = load_boundary_config()
+
+    source_configs = _source_configs_with_defaults(config_payload)
+    missing = [source for source in BOUNDARY_SOURCE_KINDS if source not in source_configs]
+    if missing:
+        raise EarthBoundaryBuildError(
+            f"Missing Earth boundary source configs: {', '.join(missing)}",
+            code="missing_sources",
+            details={"missing": missing},
+        )
+
+    sources = []
+    source_keys = list(BOUNDARY_SOURCE_KINDS)
+    for index, source_key in enumerate(source_keys):
+        source_config = source_configs[source_key]
+        if not isinstance(source_config, dict):
+            raise EarthBoundaryBuildError(
+                f"{source_key} config must be an object",
+                code="invalid_config",
+                details={"source": source_key},
+            )
+        source_start = 8 + index * 18
+        source_end = source_start + 18
+        if progress_callback:
+            progress_callback(source_start, "download", f"正在下载 {source_key}")
+
+        def report_download_progress(ratio: float | None, downloaded: int, total: int) -> None:
+            if not progress_callback:
+                return
+            if ratio is None:
+                progress_callback(source_start + 8, "download", f"{source_key} 已下载 {downloaded} bytes")
+                return
+            progress_callback(
+                source_start + (source_end - source_start) * ratio,
+                "download",
+                f"{source_key} 下载 {int(ratio * 100)}%",
+                downloaded_bytes=downloaded,
+                total_bytes=total,
+            )
+
+        sources.append(await _download_source(source_key, source_config, report_download_progress))
+
+    source_manifest = {
+        "schema": "planet-earth-boundary-sources/v2",
+        "sources": sources,
+        "povPolicy": _read_json(POV_POLICY_PATH),
+    }
+    if progress_callback:
+        progress_callback(65, "manifest", "正在写入边界源 manifest")
+    _write_json(SOURCE_MANIFEST_PATH, source_manifest)
+    build_input_hash = _build_input_hash(source_manifest)
+
+    boundary_manifest = _read_json(BOUNDARY_MANIFEST_PATH)
+    build_skipped = _has_current_artifacts(boundary_manifest, build_input_hash)
+    missing_tools = [tool for tool in ("tippecanoe", "pmtiles") if shutil.which(tool) is None]
+    if missing_tools and not build_skipped:
+        if progress_callback:
+            progress_callback(82, "build", "缺少 PMTiles 工具，正在生成 GeoJSON 高清包")
+        boundary_manifest = _write_high_precision_geojson_manifest(
+            sources,
+            build_input_hash,
+            missing_tools,
+        )
+        result = {
+            "status": "built_geojson_fallback",
+            "code": "missing_tools",
+            "missing_tools": missing_tools,
+            "sources": sources,
+            "boundary_manifest": _relative(BOUNDARY_MANIFEST_PATH),
+            "manifest": boundary_manifest,
+        }
+        _write_json(BUILD_RESULT_PATH, result)
+        if progress_callback:
+            progress_callback(96, "finalize", "GeoJSON 高清国界包已生成")
+        return {**get_boundary_status(), "build": result}
+
+    if build_skipped:
+        if progress_callback:
+            progress_callback(96, "unchanged", "高精国界已是最新")
+        build_result = {
+            "status": "unchanged",
+            "reason": "source manifest and build config hash unchanged",
+            "buildInputHash": build_input_hash,
+        }
+    else:
+        if progress_callback:
+            progress_callback(72, "build", "正在构建 PMTiles/MVT")
+        admin0 = next(source for source in sources if source["kind"] == "admin0-boundaries")
+        coastline = next(source for source in sources if source["kind"] == "coastline")
+        claim_lines = next(source for source in sources if source["kind"] == "claim-lines")
+        build_result = await _run_step(
+            [
+                "scripts/build_earth_boundary_pmtiles.py",
+                "--admin0-source",
+                admin0["path"],
+                "--coastline-source",
+                coastline["path"],
+                "--claims-source",
+                claim_lines["path"],
+                "--output",
+                _relative(PMTILES_ARTIFACT_PATH),
+                "--manifest",
+                _relative(BOUNDARY_MANIFEST_PATH),
+                "--build-input-hash",
+                build_input_hash,
+                "--pov-policy",
+                _relative(POV_POLICY_PATH),
+            ]
+        )
+        if progress_callback:
+            progress_callback(95, "finalize", "正在校验构建产物")
+
+    boundary_manifest = _read_json(BOUNDARY_MANIFEST_PATH)
+    boundary_stats = _directory_stats(BOUNDARY_OUTPUT_DIR)
+    result = {
+        "status": "unchanged" if build_skipped else "built",
+        "sources": sources,
+        "source_manifest": _relative(SOURCE_MANIFEST_PATH),
+        "boundary_manifest": _relative(BOUNDARY_MANIFEST_PATH),
+        "pmtiles_artifact": _relative(PMTILES_ARTIFACT_PATH),
+        "pmtiles_exists": PMTILES_ARTIFACT_PATH.exists(),
+        "boundary_stats": boundary_stats,
+        "manifest": boundary_manifest,
+        "build_result": build_result,
+    }
+    _write_json(BUILD_RESULT_PATH, result)
+    return {**get_boundary_status(), "build": result}
+
+
+async def _run_boundary_build_job(job_id: str) -> None:
+    def report(progress: float, phase: str, message: str, **extra: Any) -> None:
+        if _build_job_state.get("id") != job_id:
+            return
+        _update_job_progress(progress, phase, message, **extra)
+
+    try:
+        report(3, "prepare", "正在准备高精国界构建")
+        result = await build_boundary_assets(report)
+        _set_job_state(
+            id=job_id,
+            status="succeeded",
+            progress=100,
+            phase="complete",
+            message="高精国界构建完成",
+            finished_at=_utc_now_iso(),
+            result={
+                "provider": result.get("provider"),
+                "high_precision_ready": result.get("high_precision_ready"),
+                "pmtiles": result.get("pmtiles"),
+                "manifest": result.get("manifest"),
+            },
+        )
+        _append_job_log("高精国界构建完成")
+    except EarthBoundaryBuildError as exc:
+        _set_job_state(
+            id=job_id,
+            status="failed",
+            progress=_build_job_state.get("progress", 0),
+            phase="failed",
+            message=str(exc),
+            code=exc.code,
+            details=exc.details,
+            finished_at=_utc_now_iso(),
+        )
+        _append_job_log(str(exc))
+    except Exception as exc:  # pragma: no cover - defensive guard for background task
+        _set_job_state(
+            id=job_id,
+            status="failed",
+            progress=_build_job_state.get("progress", 0),
+            phase="failed",
+            message=str(exc),
+            code="build_failed",
+            finished_at=_utc_now_iso(),
+        )
+        _append_job_log(str(exc))
+
+
+async def start_boundary_build_job() -> dict[str, Any]:
+    global _build_task
+    async with _build_job_lock:
+        if _build_task and not _build_task.done():
+            return {"accepted": False, "job": _public_job_state()}
+        job_id = uuid4().hex
+        _set_job_state(
+            id=job_id,
+            status="queued",
+            progress=0,
+            phase="queued",
+            message="高精国界构建已加入队列",
+            logs=[],
+            started_at=_utc_now_iso(),
+            finished_at=None,
+            code=None,
+            details=None,
+        )
+        _append_job_log("高精国界构建已启动")
+        _build_task = asyncio.create_task(_run_boundary_build_job(job_id))
+        return {"accepted": True, "job": _public_job_state()}
--- a/backend/app/services/earth_news.py
+++ b/backend/app/services/earth_news.py
@@ -1,11 +1,13 @@
 from __future__ import annotations

 import asyncio
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from datetime import UTC, datetime
 from email.utils import parsedate_to_datetime
 import hashlib
 import html
+import json
+import math
 import re
 from typing import Any
 from urllib.parse import quote
@@ -13,6 +15,13 @@ import xml.etree.ElementTree as ET

 import httpx
 from bs4 import BeautifulSoup
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.core.countries import COUNTRY_VARIANTS_MAP, get_country_centroid, normalize_country
+from app.ai_tasks.prompts import EffectiveAIPrompt, get_effective_prompt
+from app.schemas.ai import SituationalAnalysisRequest
+from app.services.ai_client import AIProviderClient
+from app.services.location.resolvers.nominatim import build_default_nominatim_geocoder


 USER_AGENT = "PlanetEarthNewsBoard/1.0 (+https://planet.local)"
@@ -20,6 +29,11 @@ REQUEST_TIMEOUT = 12.0
 MAX_ITEMS_PER_SOURCE = 6
 MAX_ITEMS_TOTAL = 12
 STALE_CACHE_MAX_AGE_SECONDS = 60 * 45
+RSS_SUPPLEMENT_MAX_AGE_SECONDS = STALE_CACHE_MAX_AGE_SECONDS
+MAX_TARGET_INFERENCE_CONCURRENCY = 3
+TARGET_INFERENCE_TIMEOUT_SECONDS = 6.0
+DEFAULT_NEWS_LOCALE = "zh-CN"
+NEWS_ENRICH_PROMPT_KEY = "earth.news.enrich"


@dataclass(frozen=True)
@@ -49,6 +63,17 @@ class NewsFeedSource:
    priority: int = 100


+@dataclass(frozen=True)
+class NewsTargetLocation:
+    latitude: float
+    longitude: float
+    label: str
+    source: str
+    confidence: float | None = None
+    country: str | None = None
+    city: str | None = None
+
+
@dataclass
 class ParsedNewsItem:
    id: str
@@ -60,6 +85,18 @@ class ParsedNewsItem:
    feed_region: str
    homepage_url: str
    published_at: datetime | None
+    content_language: str = "en"
+    localizations: dict[str, dict[str, str]] = field(default_factory=dict)
+    enrichment_status: str = "pending"
+    enrichment_error: str | None = None
+    enriched_at: datetime | None = None
+    target_location: NewsTargetLocation | None = None
+    target_resolution_stage: str = "unresolved"
+    target_ai_attempted: bool = False
+    target_ai_status: str = "not_attempted"
+    target_ai_error: str | None = None
+    target_debug_note: str | None = None
+    location_patch: dict[str, Any] | None = None


@dataclass
@@ -236,6 +273,17 @@ NEWS_FEED_SOURCES: tuple[NewsFeedSource, ...] = (


 _REGION_CACHE: dict[str, CachedRegionFeed] = {}
+_news_target_geocode = build_default_nominatim_geocoder(user_agent=USER_AGENT)
+_CITY_HINTS: tuple[dict[str, str | None], ...] = (
+    {"name": "Beijing", "country": "中国"},
+    {"name": "Havana", "country": "古巴"},
+    {"name": "Kyiv", "country": "乌克兰"},
+    {"name": "Bangkok", "country": "泰国"},
+    {"name": "Tehran", "country": "伊朗"},
+    {"name": "Moscow", "country": "俄罗斯"},
+    {"name": "Taipei", "country": "中国（台湾）"},
+    {"name": "Hong Kong", "country": "中国（香港）"},
+)


 def determine_focus_region(lat: float | None, lon: float | None) -> str:
@@ -258,6 +306,462 @@ def get_region_anchor(region: str) -> RegionAnchor:
    return REGION_ANCHORS.get(region, REGION_ANCHORS["global"])


+def _coerce_str(value: Any) -> str | None:
+    if value is None:
+        return None
+    if not isinstance(value, str):
+        value = str(value)
+    cleaned = re.sub(r"\s+", " ", value).strip()
+    return cleaned or None
+
+
+def _contains_location_alias(text: str, alias: str) -> bool:
+    normalized_alias = _coerce_str(alias)
+    if not normalized_alias:
+        return False
+    if re.search(r"[A-Za-z]", normalized_alias):
+        pattern = r"(?<![A-Za-z])" + re.escape(normalized_alias) + r"(?![A-Za-z])"
+        return re.search(pattern, text, flags=re.IGNORECASE) is not None
+    return normalized_alias in text
+
+
+def _iter_searchable_country_variants(
+    canonical: str,
+    variants: list[str],
+) -> tuple[str, ...]:
+    searchable: list[str] = []
+    seen: set[str] = set()
+    for variant in (canonical, *variants):
+        normalized = _coerce_str(variant)
+        if not normalized:
+            continue
+        if re.fullmatch(r"[A-Z]{2,3}", normalized):
+            continue
+        if len(normalized) <= 2:
+            continue
+        key = normalized.casefold()
+        if key in seen:
+            continue
+        seen.add(key)
+        searchable.append(normalized)
+    return tuple(searchable)
+
+
+def _coerce_float(value: Any) -> float | None:
+    try:
+        parsed = float(value)
+    except (TypeError, ValueError):
+        return None
+    if not math.isfinite(parsed):
+        return None
+    return parsed
+
+
+def _first_json_object(text: str) -> dict[str, Any] | None:
+    if not text:
+        return None
+    decoder = json.JSONDecoder()
+    for index, char in enumerate(text):
+        if char != "{":
+            continue
+        try:
+            payload, _ = decoder.raw_decode(text[index:])
+        except ValueError:
+            continue
+        if isinstance(payload, dict):
+            return payload
+    return None
+
+
+def _normalize_localizations(value: Any) -> dict[str, dict[str, str]]:
+    if not isinstance(value, dict):
+        return {}
+    normalized: dict[str, dict[str, str]] = {}
+    for locale, payload in value.items():
+        locale_key = _coerce_str(locale)
+        if not locale_key or not isinstance(payload, dict):
+            continue
+        title = _coerce_str(payload.get("title"))
+        summary = _coerce_str(payload.get("summary"))
+        entry: dict[str, str] = {}
+        if title:
+            entry["title"] = title
+        if summary:
+            entry["summary"] = summary
+        if entry:
+            normalized[locale_key] = entry
+    return normalized
+
+
+def _get_locale_text(
+    item: ParsedNewsItem,
+    key: str,
+    *,
+    locale: str = DEFAULT_NEWS_LOCALE,
+) -> str:
+    localized = item.localizations.get(locale)
+    if isinstance(localized, dict):
+        value = _coerce_str(localized.get(key))
+        if value:
+            return value
+    return ""
+
+
+def _has_default_localization(item: ParsedNewsItem) -> bool:
+    localized = item.localizations.get(DEFAULT_NEWS_LOCALE)
+    if not isinstance(localized, dict):
+        return False
+    return bool(_coerce_str(localized.get("title")) and _coerce_str(localized.get("summary")))
+
+
+def apply_enrichment_patch_to_item(
+    item: ParsedNewsItem,
+    patch: dict[str, Any],
+) -> ParsedNewsItem:
+    item.location_patch = patch
+    if "content_language" in patch:
+        item.content_language = _coerce_str(patch.get("content_language")) or item.content_language
+    if "localizations" in patch:
+        item.localizations = _normalize_localizations(patch.get("localizations"))
+    if "enrichment_status" in patch:
+        item.enrichment_status = _coerce_str(patch.get("enrichment_status")) or item.enrichment_status
+    if "enrichment_error" in patch:
+        item.enrichment_error = _coerce_str(patch.get("enrichment_error"))
+    if "enriched_at" in patch:
+        item.enriched_at = _parse_datetime(_coerce_str(patch.get("enriched_at")))
+    return item
+
+
+async def _geocode_target_location(query: str) -> dict[str, Any] | None:
+    return await asyncio.to_thread(_news_target_geocode, query)
+
+
+async def _build_target_location_from_payload(
+    payload: dict[str, Any],
+) -> NewsTargetLocation | None:
+    country = normalize_country(payload.get("country"))
+    city = _coerce_str(payload.get("city"))
+    matched_location_name = _coerce_str(payload.get("matched_location_name"))
+    confidence = _coerce_float(payload.get("confidence"))
+    if confidence is not None:
+        confidence = max(0.0, min(confidence, 1.0))
+
+    latitude = _coerce_float(payload.get("latitude"))
+    longitude = _coerce_float(payload.get("longitude"))
+    if latitude not in (None, 0.0) and longitude not in (None, 0.0):
+        label = matched_location_name or ", ".join(part for part in (city, country) if part) or "关联位置"
+        return NewsTargetLocation(
+            latitude=latitude,
+            longitude=longitude,
+            label=label,
+            source="ai_inferred_target",
+            confidence=confidence,
+            country=country,
+            city=city,
+        )
+
+    geocode_queries: list[str] = []
+    for value in (
+        ", ".join(part for part in (city, country) if part),
+        matched_location_name,
+        city,
+        country,
+    ):
+        normalized = _coerce_str(value)
+        if normalized and normalized not in geocode_queries:
+            geocode_queries.append(normalized)
+
+    for query in geocode_queries:
+        try:
+            result = await _geocode_target_location(query)
+        except Exception:
+            continue
+        if not isinstance(result, dict):
+            continue
+        latitude = _coerce_float(result.get("lat"))
+        longitude = _coerce_float(result.get("lon"))
+        if latitude in (None, 0.0) or longitude in (None, 0.0):
+            continue
+        label = (
+            _coerce_str(result.get("display_name"))
+            or matched_location_name
+            or ", ".join(part for part in (city, country) if part)
+            or query
+        )
+        return NewsTargetLocation(
+            latitude=latitude,
+            longitude=longitude,
+            label=label,
+            source="ai_inferred_target",
+            confidence=confidence,
+            country=country,
+            city=city,
+        )
+
+    centroid = get_country_centroid(country)
+    if centroid:
+        label = matched_location_name or city or country or "关联位置"
+        return NewsTargetLocation(
+            latitude=centroid["latitude"],
+            longitude=centroid["longitude"],
+            label=label,
+            source="ai_inferred_target",
+            confidence=confidence,
+            country=country,
+            city=city,
+        )
+
+    return None
+
+
+async def _extract_target_location_from_text(item: ParsedNewsItem) -> NewsTargetLocation | None:
+    combined_text = " ".join(part for part in (item.title, item.summary) if part).strip()
+    if not combined_text:
+        return None
+
+    for hint in _CITY_HINTS:
+        city_name = _coerce_str(hint.get("name"))
+        if not city_name or not _contains_location_alias(combined_text, city_name):
+            continue
+        country = normalize_country(hint.get("country"))
+        geocode_query = ", ".join(part for part in (city_name, country) if part)
+        try:
+            result = await _geocode_target_location(geocode_query)
+        except Exception:
+            result = None
+        if isinstance(result, dict):
+            latitude = _coerce_float(result.get("lat"))
+            longitude = _coerce_float(result.get("lon"))
+            if latitude not in (None, 0.0) and longitude not in (None, 0.0):
+                return NewsTargetLocation(
+                    latitude=latitude,
+                    longitude=longitude,
+                    label=_coerce_str(result.get("display_name")) or geocode_query,
+                    source="headline_location_hint",
+                    confidence=0.78,
+                    country=country,
+                    city=city_name,
+                )
+        centroid = get_country_centroid(country)
+        if centroid:
+            return NewsTargetLocation(
+                latitude=centroid["latitude"],
+                longitude=centroid["longitude"],
+                label=geocode_query,
+                source="headline_location_hint",
+                confidence=0.68,
+                country=country,
+                city=city_name,
+            )
+
+    for canonical, variants in COUNTRY_VARIANTS_MAP.items():
+        if not get_country_centroid(canonical):
+            continue
+        searchable_variants = _iter_searchable_country_variants(canonical, variants)
+        if not any(_contains_location_alias(combined_text, variant) for variant in searchable_variants):
+            continue
+        centroid = get_country_centroid(canonical)
+        if not centroid:
+            continue
+        return NewsTargetLocation(
+            latitude=centroid["latitude"],
+            longitude=centroid["longitude"],
+            label=canonical,
+            source="headline_country_hint",
+            confidence=0.62,
+            country=canonical,
+            city=None,
+        )
+
+    return None
+
+
+async def _infer_news_target_location(
+    item: ParsedNewsItem,
+    *,
+    provider_client: AIProviderClient | None,
+    prompt: EffectiveAIPrompt | None = None,
+) -> NewsTargetLocation | None:
+    target, _localizations = await _infer_news_enrichment(
+        item,
+        provider_client=provider_client,
+        prompt=prompt,
+    )
+    return target
+
+
+async def _infer_news_enrichment(
+    item: ParsedNewsItem,
+    *,
+    provider_client: AIProviderClient | None,
+    prompt: EffectiveAIPrompt | None = None,
+) -> tuple[NewsTargetLocation | None, dict[str, dict[str, str]]]:
+    text_hint = await _extract_target_location_from_text(item)
+    content_error: str | None = None
+    if text_hint is not None and text_hint.city:
+        item.target_resolution_stage = text_hint.source
+        item.target_ai_attempted = False
+        item.target_ai_status = "skipped_text_hint"
+        item.target_ai_error = None
+        item.target_debug_note = f"text hint matched {text_hint.label}"
+    localizations: dict[str, dict[str, str]] = {}
+
+    if provider_client is None:
+        if text_hint is None or not text_hint.city:
+            item.target_resolution_stage = text_hint.source if text_hint is not None else "unresolved"
+            item.target_ai_attempted = False
+            item.target_ai_status = "unavailable"
+            item.target_ai_error = "AI provider is not configured or unavailable for earth-feed."
+            item.target_debug_note = (
+                f"text hint matched {text_hint.label}" if text_hint is not None else "no text location hint matched"
+            )
+        item.enrichment_status = "unavailable"
+        item.enrichment_error = "AI provider is not configured or unavailable for earth-feed."
+        return text_hint, localizations
+
+    if text_hint is None or not text_hint.city:
+        item.target_ai_attempted = True
+        item.target_ai_status = "attempted"
+        item.target_ai_error = None
+        item.target_debug_note = (
+            f"text hint matched {text_hint.label}" if text_hint is not None else "no text location hint matched"
+        )
+    item.enrichment_status = "attempted"
+    item.enrichment_error = None
+
+    prompt = prompt or await get_effective_prompt(None, NEWS_ENRICH_PROMPT_KEY)
+    request = SituationalAnalysisRequest(
+        title="Enrich Earth news item with event location and zh-CN content",
+        objective=prompt.prompt,
+        system_prompt=prompt.system_prompt or None,
+        context={
+            "news_item": {
+                "title": item.title,
+                "summary": item.summary,
+                "source": item.source,
+                "feed_name": item.feed_name,
+                "feed_region": item.feed_region,
+                "url": item.url,
+                "published_at": (
+                    item.published_at.isoformat().replace("+00:00", "Z")
+                    if item.published_at
+                    else None
+                ),
+            },
+            "required_json_schema": {
+                "location": {
+                    "country": "string|null",
+                    "city": "string|null",
+                    "matched_location_name": "string|null",
+                    "latitude": "number|null",
+                    "longitude": "number|null",
+                    "confidence": "number from 0 to 1",
+                    "reasoning_summary": "short string",
+                },
+                "localizations": {
+                    "zh-CN": {
+                        "title": "faithful Simplified Chinese title",
+                        "summary": "one-sentence newswire-style Simplified Chinese lead summary",
+                    }
+                },
+            },
+        },
+        constraints=[
+            "Return only strict JSON. Do not wrap it in markdown.",
+            "For localizations, do not add facts that are absent from the RSS headline, description, source, or date.",
+            "Write zh-CN summary as one concise newswire-style sentence, like a breaking-news lead.",
+            "If the RSS description is thin, write a conservative one-sentence summary that says only what is supported.",
+            "Keep zh-CN summary factual, non-promotional, and avoid colon-heavy keyword labels.",
+            "Prefer the event location, not the newsroom or publisher headquarters.",
+            "When a country visit or summit is the clear topic but the city is omitted, use the most likely host city only if it is broadly public knowledge.",
+            "Use null for unknown fields instead of inventing details.",
+            "Calibrate confidence conservatively: 0.75+ only when the city is strongly supported, 0.55-0.74 for country-level or likely city inference, below 0.55 when weak.",
+        ],
+    )
+    try:
+        response = await provider_client.analyze(request)
+    except Exception as exc:
+        if text_hint is None or not text_hint.city:
+            item.target_resolution_stage = text_hint.source if text_hint is not None else "unresolved"
+            item.target_ai_status = "provider_error"
+            item.target_ai_error = str(exc)
+        item.enrichment_status = "provider_error"
+        item.enrichment_error = str(exc)
+        return text_hint, localizations
+
+    payload = _first_json_object(response.content)
+    if not isinstance(payload, dict):
+        if text_hint is None or not text_hint.city:
+            item.target_resolution_stage = text_hint.source if text_hint is not None else "unresolved"
+            item.target_ai_status = "parse_error"
+            item.target_ai_error = "AI response did not contain a parseable JSON object."
+        item.enrichment_status = "parse_error"
+        item.enrichment_error = "AI response did not contain a parseable JSON object."
+        return text_hint, localizations
+
+    localizations = _normalize_localizations(payload.get("localizations"))
+    if not localizations:
+        content_error = "AI returned no usable localizations."
+
+    location_payload = payload.get("location") if isinstance(payload.get("location"), dict) else payload
+    if text_hint is not None and text_hint.city:
+        target = text_hint
+    else:
+        target = await _build_target_location_from_payload(location_payload)
+    if target is None:
+        item.target_resolution_stage = text_hint.source if text_hint is not None else "unresolved"
+        item.target_ai_status = "no_result"
+        item.target_ai_error = "AI returned no usable target coordinates or geocodeable location."
+        target = text_hint
+    elif target.confidence is not None and target.confidence < 0.45:
+        item.target_resolution_stage = text_hint.source if text_hint is not None else "unresolved"
+        item.target_ai_status = "low_confidence"
+        item.target_ai_error = f"AI target confidence too low: {target.confidence:.2f}"
+        target = text_hint
+    else:
+        item.target_resolution_stage = target.source
+        item.target_ai_status = "success"
+        item.target_ai_error = None
+        item.target_debug_note = f"ai inferred {target.label}"
+
+    item.localizations = localizations
+    if localizations and item.target_ai_status in {"success", "skipped_text_hint"}:
+        item.enrichment_status = "success"
+        item.enrichment_error = None
+    elif localizations:
+        item.enrichment_status = "content_only"
+        item.enrichment_error = item.target_ai_error
+    else:
+        item.enrichment_status = "location_only" if target is not None else "no_result"
+        item.enrichment_error = content_error or item.target_ai_error
+    item.enriched_at = datetime.now(UTC) if localizations else None
+    return target, localizations
+
+
+async def _enrich_items_with_target_locations(
+    items: list[ParsedNewsItem],
+    *,
+    provider_client: AIProviderClient | None,
+    prompt: EffectiveAIPrompt | None = None,
+) -> list[ParsedNewsItem]:
+    if not items:
+        return items
+
+    semaphore = asyncio.Semaphore(MAX_TARGET_INFERENCE_CONCURRENCY)
+
+    async def enrich(item: ParsedNewsItem) -> ParsedNewsItem:
+        async with semaphore:
+            target = await _infer_news_target_location(
+                item,
+                provider_client=provider_client,
+                prompt=prompt,
+            )
+        item.target_location = target
+        return item
+
+    return list(await asyncio.gather(*(enrich(item) for item in items)))
+
+
 def get_sources_for_region(region: str) -> list[NewsFeedSource]:
    return sorted(
        [source for source in NEWS_FEED_SOURCES if source.region in {"global", region}],
@@ -385,23 +889,170 @@ def _serialize_sources(sources: list[NewsFeedSource]) -> list[dict[str, Any]]:
    ]


-def _serialize_item(item: ParsedNewsItem, *, active_region: str) -> dict[str, Any]:
-    published_at = item.published_at
+def _serialize_anchor(anchor: RegionAnchor) -> dict[str, Any]:
+    return {
+        "region": anchor.region,
+        "label": anchor.label,
+        "latitude": anchor.latitude,
+        "longitude": anchor.longitude,
+    }
+
+
+def _serialize_target(target: NewsTargetLocation | None) -> dict[str, Any] | None:
+    if target is None:
+        return None
+    return {
+        "latitude": target.latitude,
+        "longitude": target.longitude,
+        "label": target.label,
+        "source": target.source,
+        "confidence": target.confidence,
+        "country": target.country,
+        "city": target.city,
+    }
+
+
+def _serialize_enriched_at(value: datetime | None) -> str | None:
+    return value.isoformat().replace("+00:00", "Z") if value else None
+
+
+def _content_patch(item: ParsedNewsItem) -> dict[str, Any]:
+    return {
+        "content_language": item.content_language,
+        "localizations": item.localizations,
+        "enrichment_status": item.enrichment_status,
+        "enrichment_error": item.enrichment_error,
+        "enriched_at": _serialize_enriched_at(item.enriched_at),
+    }
+
+
+def build_anchor_location_patch(
+    item: ParsedNewsItem,
+    *,
+    queued: bool = False,
+    queue_available: bool | None = None,
+) -> dict[str, Any]:
    anchor = get_region_anchor(item.feed_region)
+    if queued:
+        resolution_stage = "queued"
+        ai_status = "queued"
+        debug_note = "queued for async target location inference"
+    else:
+        resolution_stage = item.target_resolution_stage
+        ai_status = item.target_ai_status
+        debug_note = item.target_debug_note
+    content_patch = _content_patch(item)
+    if queued and content_patch["enrichment_status"] == "pending":
+        content_patch["enrichment_status"] = "queued"
+    return {
+        "latitude": anchor.latitude,
+        "longitude": anchor.longitude,
+        "location_label": anchor.label,
+        "location_source": "region_anchor",
+        "verified": False,
+        "location_meta": {
+            "resolution_stage": resolution_stage,
+            "ai_attempted": item.target_ai_attempted,
+            "ai_status": ai_status,
+            "ai_error": item.target_ai_error,
+            "debug_note": debug_note,
+            "queue_available": queue_available,
+            "target": None,
+            "anchor": _serialize_anchor(anchor),
+        },
+        **content_patch,
+    }
+
+
+def build_target_location_patch(item: ParsedNewsItem, target: NewsTargetLocation | None) -> dict[str, Any]:
+    if target is None:
+        return build_anchor_location_patch(item)
+    anchor = get_region_anchor(item.feed_region)
+    return {
+        "latitude": target.latitude,
+        "longitude": target.longitude,
+        "location_label": target.label,
+        "location_source": target.source,
+        "verified": True,
+        "location_meta": {
+            "resolution_stage": item.target_resolution_stage,
+            "ai_attempted": item.target_ai_attempted,
+            "ai_status": item.target_ai_status,
+            "ai_error": item.target_ai_error,
+            "debug_note": item.target_debug_note,
+            "target": _serialize_target(target),
+            "anchor": _serialize_anchor(anchor),
+        },
+        **_content_patch(item),
+    }
+
+
+def build_target_location_job_payload(item: ParsedNewsItem) -> dict[str, Any]:
+    published_at = item.published_at
    return {
        "id": item.id,
        "title": item.title,
        "summary": item.summary,
+        "content_language": item.content_language,
+        "localizations": item.localizations,
+        "enrichment_status": item.enrichment_status,
+        "enrichment_error": item.enrichment_error,
+        "enriched_at": _serialize_enriched_at(item.enriched_at),
+        "url": item.url,
+        "source": item.source,
+        "feed_name": item.feed_name,
+        "feed_region": item.feed_region,
+        "homepage_url": item.homepage_url,
+        "published_at": published_at.isoformat().replace("+00:00", "Z") if published_at else None,
+    }
+
+
+def parsed_news_item_from_job_payload(payload: dict[str, Any]) -> ParsedNewsItem:
+    return ParsedNewsItem(
+        id=str(payload.get("id") or ""),
+        title=str(payload.get("title") or ""),
+        summary=str(payload.get("summary") or ""),
+        content_language=str(payload.get("content_language") or "en"),
+        localizations=_normalize_localizations(payload.get("localizations")),
+        enrichment_status=str(payload.get("enrichment_status") or "pending"),
+        enrichment_error=_coerce_str(payload.get("enrichment_error")),
+        enriched_at=_parse_datetime(_coerce_str(payload.get("enriched_at"))),
+        url=str(payload.get("url") or ""),
+        source=str(payload.get("source") or ""),
+        feed_name=str(payload.get("feed_name") or ""),
+        feed_region=str(payload.get("feed_region") or "global"),
+        homepage_url=str(payload.get("homepage_url") or ""),
+        published_at=_parse_datetime(_coerce_str(payload.get("published_at"))),
+    )
+
+
+def _serialize_item(item: ParsedNewsItem, *, active_region: str) -> dict[str, Any]:
+    published_at = item.published_at
+    location_patch = item.location_patch or build_target_location_patch(item, item.target_location)
+    return {
+        "id": item.id,
+        "title": item.title,
+        "summary": item.summary,
+        "content_language": item.content_language,
+        "localizations": item.localizations,
+        "display_title": _get_locale_text(item, "title"),
+        "display_summary": _get_locale_text(item, "summary"),
        "url": item.url,
        "source": item.source,
        "feed_name": item.feed_name,
        "region": item.feed_region,
+        "display_region": get_region_anchor(item.feed_region).label,
        "homepage_url": item.homepage_url,
        "published_at": published_at.isoformat().replace("+00:00", "Z") if published_at else None,
-        "latitude": anchor.latitude,
-        "longitude": anchor.longitude,
-        "location_label": anchor.label,
-        "location_inferred": True,
+        "latitude": location_patch["latitude"],
+        "longitude": location_patch["longitude"],
+        "location_label": location_patch["location_label"],
+        "location_source": location_patch["location_source"],
+        "verified": location_patch["verified"],
+        "location_meta": location_patch["location_meta"],
+        "enrichment_status": item.enrichment_status,
+        "enrichment_error": item.enrichment_error,
+        "enriched_at": _serialize_enriched_at(item.enriched_at),
        "is_focus_match": item.feed_region == active_region,
    }

@@ -426,6 +1077,7 @@ def _build_payload(
            "lon": lon,
            "region": active_region,
            "label": profile.label,
+            "display_region": get_region_anchor(active_region).label,
            "accent": profile.accent,
        },
        "sources": _serialize_sources(sources),
@@ -472,6 +1124,68 @@ def _store_region_cache(region: str, *, items: list[ParsedNewsItem], sources: li
    )


+async def _apply_cached_locations_and_enqueue(items: list[ParsedNewsItem]) -> list[ParsedNewsItem]:
+    if not items:
+        return items
+
+    from app.services.earth_news_queue import (
+        enqueue_target_location_job,
+        get_cached_target_location_patch,
+    )
+
+    async def enqueue_item(item: ParsedNewsItem, *, force: bool = False) -> bool:
+        return await enqueue_target_location_job(build_target_location_job_payload(item), force=force)
+
+    async def apply_location(item: ParsedNewsItem) -> ParsedNewsItem:
+        cached_patch = await get_cached_target_location_patch(item.id)
+        if cached_patch:
+            apply_enrichment_patch_to_item(item, cached_patch)
+            if not _has_default_localization(item):
+                queued = await enqueue_item(item, force=True)
+                if queued and item.enrichment_status in {
+                    "pending",
+                    "unavailable",
+                    "provider_error",
+                    "parse_error",
+                    "no_result",
+                    "location_only",
+                }:
+                    item.enrichment_status = "queued"
+            return item
+
+        queued = await enqueue_item(item)
+        item.location_patch = build_anchor_location_patch(
+            item,
+            queued=queued,
+            queue_available=queued,
+        )
+        return item
+
+    return list(await asyncio.gather(*(apply_location(item) for item in items)))
+
+
+async def _enqueue_unverified_locations(items: list[ParsedNewsItem]) -> None:
+    if not items:
+        return
+
+    from app.services.earth_news_queue import enqueue_target_location_job
+
+    await asyncio.gather(
+        *(
+            enqueue_target_location_job(
+                build_target_location_job_payload(item),
+                force=not _has_default_localization(item),
+            )
+            for item in items
+            if (
+                item.location_patch is None
+                or item.location_patch.get("verified") is False
+                or not _has_default_localization(item)
+            )
+        )
+    )
+
+
 async def _fetch_source(
    client: httpx.AsyncClient,
    source: NewsFeedSource,
@@ -484,11 +1198,10 @@ async def _fetch_source(
        return source, [], str(exc)


-async def get_earth_news_payload(lat: float | None = None, lon: float | None = None) -> dict[str, Any]:
-    active_region = determine_focus_region(lat, lon)
-    sources = get_sources_for_region(active_region)
+async def _fetch_rss_items_for_sources(
+    sources: list[NewsFeedSource],
+) -> tuple[list[ParsedNewsItem], list[str]]:
    errors: list[str] = []
-
    async with httpx.AsyncClient(
        timeout=REQUEST_TIMEOUT,
        follow_redirects=True,
@@ -502,9 +1215,29 @@ async def get_earth_news_payload(lat: float | None = None, lon: float | None = N
            errors.append(f"{source.name}: {error}")
            continue
        fetched_items.extend(items)
+    return fetched_items, errors

+
+def _needs_rss_supplement(*, item_count: int, newest_at: datetime | None) -> bool:
+    if item_count < MAX_ITEMS_TOTAL:
+        return True
+    if newest_at is None:
+        return True
+    age_seconds = (datetime.now(UTC) - newest_at).total_seconds()
+    return age_seconds > RSS_SUPPLEMENT_MAX_AGE_SECONDS
+
+
+async def _get_earth_news_payload_from_rss_only(
+    *,
+    lat: float | None,
+    lon: float | None,
+    active_region: str,
+    sources: list[NewsFeedSource],
+) -> dict[str, Any]:
+    fetched_items, errors = await _fetch_rss_items_for_sources(sources)
    ranked_items = _rank_and_trim_items(fetched_items, active_region=active_region)
    if ranked_items:
+        ranked_items = await _apply_cached_locations_and_enqueue(ranked_items)
        _store_region_cache(active_region, items=ranked_items, sources=sources)
        return _build_payload(
            lat=lat,
@@ -518,6 +1251,7 @@ async def get_earth_news_payload(lat: float | None = None, lon: float | None = N

    cached = _get_cached_region_feed(active_region)
    if cached:
+        cached.items = await _apply_cached_locations_and_enqueue(cached.items)
        return _build_payload(
            lat=lat,
            lon=lon,
@@ -538,3 +1272,55 @@ async def get_earth_news_payload(lat: float | None = None, lon: float | None = N
        errors=errors,
        stale=False,
    )
+
+
+async def get_earth_news_payload(
+    lat: float | None = None,
+    lon: float | None = None,
+    *,
+    provider_client: AIProviderClient | None = None,
+    db: AsyncSession | None = None,
+) -> dict[str, Any]:
+    del provider_client
+    active_region = determine_focus_region(lat, lon)
+    sources = get_sources_for_region(active_region)
+
+    if db is None:
+        return await _get_earth_news_payload_from_rss_only(
+            lat=lat,
+            lon=lon,
+            active_region=active_region,
+            sources=sources,
+        )
+
+    from app.services.earth_news_store import (
+        get_earth_news_freshness,
+        list_earth_news_items,
+        upsert_earth_news_items,
+    )
+
+    errors: list[str] = []
+    item_count, newest_at = await get_earth_news_freshness(db, active_region=active_region)
+    should_supplement = _needs_rss_supplement(item_count=item_count, newest_at=newest_at)
+    if should_supplement:
+        fetched_items, errors = await _fetch_rss_items_for_sources(sources)
+        ranked_fetched_items = _rank_and_trim_items(fetched_items, active_region=active_region)
+        await upsert_earth_news_items(db, ranked_fetched_items)
+
+    items = await list_earth_news_items(
+        db,
+        active_region=active_region,
+        limit=MAX_ITEMS_TOTAL,
+    )
+    await _enqueue_unverified_locations(items)
+    stale = bool(errors and items)
+
+    return _build_payload(
+        lat=lat,
+        lon=lon,
+        active_region=active_region,
+        items=items,
+        sources=sources,
+        errors=errors,
+        stale=stale,
+    )
--- a/backend/app/services/earth_news_queue.py
+++ b/backend/app/services/earth_news_queue.py
@@ -0,0 +1,234 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+import json
+from typing import Any, Protocol
+
+import redis.asyncio as redis
+from redis.exceptions import ResponseError
+
+from app.core.config import settings
+from app.core.logging import get_logger
+
+
+logger = get_logger(__name__, service="earth_news")
+
+TARGET_LOCATION_STREAM = "earth_news:target_location:jobs"
+TARGET_LOCATION_GROUP = "earth_news_target_location"
+TARGET_LOCATION_DEAD_LETTER_STREAM = "earth_news:target_location:dead"
+TARGET_LOCATION_RESULT_TTL_SECONDS = 60 * 60 * 12
+TARGET_LOCATION_JOB_DEDUP_TTL_SECONDS = 60 * 60 * 6
+TARGET_LOCATION_MAX_ATTEMPTS = 3
+
+_redis_client: redis.Redis | None = None
+
+
+@dataclass(frozen=True)
+class NewsTargetLocationMessage:
+    message_id: str
+    item_id: str
+    payload: dict[str, Any]
+    attempts: int = 0
+
+
+class NewsTargetLocationQueue(Protocol):
+    async def enqueue(self, *, item_id: str, payload: dict[str, Any], force: bool = False) -> bool:
+        ...
+
+    async def consume_batch(
+        self,
+        *,
+        consumer_name: str,
+        count: int,
+        block_ms: int,
+    ) -> list[NewsTargetLocationMessage]:
+        ...
+
+    async def ack(self, message_id: str) -> None:
+        ...
+
+    async def retry_or_dead_letter(
+        self,
+        message: NewsTargetLocationMessage,
+        *,
+        error: str,
+    ) -> None:
+        ...
+
+
+def _get_redis_client() -> redis.Redis:
+    global _redis_client
+    if _redis_client is None:
+        _redis_client = redis.from_url(settings.REDIS_URL, decode_responses=True)
+    return _redis_client
+
+
+def _result_key(item_id: str) -> str:
+    return f"earth_news:target_location:result:{item_id}"
+
+
+def _queued_key(item_id: str) -> str:
+    return f"earth_news:target_location:queued:{item_id}"
+
+
+class RedisStreamsNewsTargetLocationQueue:
+    def __init__(self, client: redis.Redis | None = None) -> None:
+        self.client = client or _get_redis_client()
+        self._group_ready = False
+
+    async def _ensure_group(self) -> None:
+        if self._group_ready:
+            return
+        try:
+            await self.client.xgroup_create(
+                TARGET_LOCATION_STREAM,
+                TARGET_LOCATION_GROUP,
+                id="0",
+                mkstream=True,
+            )
+        except ResponseError as exc:
+            if "BUSYGROUP" not in str(exc):
+                raise
+        self._group_ready = True
+
+    async def enqueue(self, *, item_id: str, payload: dict[str, Any], force: bool = False) -> bool:
+        await self._ensure_group()
+        if force:
+            await self.client.delete(_result_key(item_id), _queued_key(item_id))
+        elif await self.client.exists(_result_key(item_id)):
+            return False
+        queued = await self.client.set(
+            _queued_key(item_id),
+            "1",
+            nx=True,
+            ex=TARGET_LOCATION_JOB_DEDUP_TTL_SECONDS,
+        )
+        if not queued:
+            return bool(await self.client.exists(_queued_key(item_id)))
+        await self.client.xadd(
+            TARGET_LOCATION_STREAM,
+            {
+                "item_id": item_id,
+                "attempts": "0",
+                "payload": json.dumps(payload, ensure_ascii=False),
+            },
+        )
+        return True
+
+    async def consume_batch(
+        self,
+        *,
+        consumer_name: str,
+        count: int,
+        block_ms: int,
+    ) -> list[NewsTargetLocationMessage]:
+        await self._ensure_group()
+        streams = await self.client.xreadgroup(
+            TARGET_LOCATION_GROUP,
+            consumer_name,
+            {TARGET_LOCATION_STREAM: ">"},
+            count=count,
+            block=block_ms,
+        )
+        messages: list[NewsTargetLocationMessage] = []
+        for _stream_name, stream_messages in streams:
+            for message_id, fields in stream_messages:
+                raw_payload = fields.get("payload")
+                item_id = fields.get("item_id")
+                if not raw_payload or not item_id:
+                    await self.ack(message_id)
+                    continue
+                try:
+                    payload = json.loads(raw_payload)
+                except json.JSONDecodeError:
+                    await self.ack(message_id)
+                    continue
+                attempts = int(fields.get("attempts") or 0)
+                messages.append(
+                    NewsTargetLocationMessage(
+                        message_id=message_id,
+                        item_id=item_id,
+                        payload=payload,
+                        attempts=attempts,
+                    )
+                )
+        return messages
+
+    async def ack(self, message_id: str) -> None:
+        await self.client.xack(TARGET_LOCATION_STREAM, TARGET_LOCATION_GROUP, message_id)
+
+    async def retry_or_dead_letter(
+        self,
+        message: NewsTargetLocationMessage,
+        *,
+        error: str,
+    ) -> None:
+        await self.ack(message.message_id)
+        if message.attempts + 1 >= TARGET_LOCATION_MAX_ATTEMPTS:
+            await self.client.xadd(
+                TARGET_LOCATION_DEAD_LETTER_STREAM,
+                {
+                    "item_id": message.item_id,
+                    "attempts": str(message.attempts + 1),
+                    "error": error,
+                    "payload": json.dumps(message.payload, ensure_ascii=False),
+                },
+            )
+            return
+        await self.client.xadd(
+            TARGET_LOCATION_STREAM,
+            {
+                "item_id": message.item_id,
+                "attempts": str(message.attempts + 1),
+                "payload": json.dumps(message.payload, ensure_ascii=False),
+            },
+        )
+
+
+def get_news_target_location_queue() -> NewsTargetLocationQueue:
+    return RedisStreamsNewsTargetLocationQueue()
+
+
+async def enqueue_target_location_job(payload: dict[str, Any], *, force: bool = False) -> bool:
+    item_id = str(payload.get("id") or "")
+    if not item_id:
+        return False
+    try:
+        queue = get_news_target_location_queue()
+        return await queue.enqueue(item_id=item_id, payload=payload, force=force)
+    except Exception as exc:
+        logger.warning_event(
+            "Failed to enqueue Earth news target location job",
+            event="earth_news.target_location.enqueue_failed",
+            context={"item_id": item_id, "error": str(exc)},
+        )
+        return False
+
+
+async def get_cached_target_location_patch(item_id: str) -> dict[str, Any] | None:
+    try:
+        raw_value = await _get_redis_client().get(_result_key(item_id))
+    except Exception as exc:
+        logger.warning_event(
+            "Failed to read Earth news target location cache",
+            event="earth_news.target_location.cache_read_failed",
+            context={"item_id": item_id, "error": str(exc)},
+        )
+        return None
+    if not raw_value:
+        return None
+    try:
+        value = json.loads(raw_value)
+    except json.JSONDecodeError:
+        return None
+    return value if isinstance(value, dict) else None
+
+
+async def save_target_location_patch(item_id: str, patch: dict[str, Any]) -> None:
+    client = _get_redis_client()
+    await client.setex(
+        _result_key(item_id),
+        TARGET_LOCATION_RESULT_TTL_SECONDS,
+        json.dumps(patch, ensure_ascii=False),
+    )
+    await client.delete(_queued_key(item_id))
--- a/backend/app/services/earth_news_store.py
+++ b/backend/app/services/earth_news_store.py
@@ -0,0 +1,244 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+from typing import Any
+
+from sqlalchemy import func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models.earth_news import EarthNewsItem
+from app.services.earth_news import (
+    ParsedNewsItem,
+    apply_enrichment_patch_to_item,
+    build_anchor_location_patch,
+)
+
+
+def _coerce_datetime(value: datetime | None) -> datetime | None:
+    if value is None:
+        return None
+    if value.tzinfo is None:
+        return value.replace(tzinfo=UTC)
+    return value.astimezone(UTC)
+
+
+def _location_patch_from_record(record: EarthNewsItem) -> dict[str, Any]:
+    return {
+        "latitude": record.latitude,
+        "longitude": record.longitude,
+        "location_label": record.location_label,
+        "location_source": record.location_source,
+        "verified": record.verified,
+        "location_meta": dict(record.location_meta or {}),
+    }
+
+
+def record_to_parsed_news_item(record: EarthNewsItem) -> ParsedNewsItem:
+    item = ParsedNewsItem(
+        id=record.id,
+        title=record.title,
+        summary=record.summary or "",
+        url=record.url,
+        source=record.source or "",
+        feed_name=record.feed_name or "",
+        feed_region=record.region or "global",
+        homepage_url=record.homepage_url or "",
+        published_at=_coerce_datetime(record.published_at),
+        content_language=record.content_language or "en",
+        localizations=dict(record.localizations or {}),
+        enrichment_status=record.enrichment_status or "pending",
+        enrichment_error=record.enrichment_error,
+        enriched_at=_coerce_datetime(record.enriched_at),
+    )
+    return apply_enrichment_patch_to_item(item, _location_patch_from_record(record))
+
+
+def _query_sort_key(active_region: str):
+    return (
+        EarthNewsItem.region != active_region,
+        EarthNewsItem.published_at.is_(None),
+        EarthNewsItem.published_at.desc().nullslast(),
+        EarthNewsItem.feed_name.asc(),
+    )
+
+
+async def list_earth_news_items(
+    db: AsyncSession,
+    *,
+    active_region: str,
+    limit: int,
+) -> list[ParsedNewsItem]:
+    regions = {"global", active_region}
+    result = await db.execute(
+        select(EarthNewsItem)
+        .where(EarthNewsItem.region.in_(regions))
+        .order_by(*_query_sort_key(active_region))
+        .limit(limit)
+    )
+    return [record_to_parsed_news_item(record) for record in result.scalars().all()]
+
+
+async def get_earth_news_freshness(
+    db: AsyncSession,
+    *,
+    active_region: str,
+) -> tuple[int, datetime | None]:
+    regions = {"global", active_region}
+    result = await db.execute(
+        select(
+            func.count(EarthNewsItem.id),
+            func.max(func.coalesce(EarthNewsItem.published_at, EarthNewsItem.last_seen_at)),
+        ).where(EarthNewsItem.region.in_(regions))
+    )
+    count, newest = result.one()
+    item_count = int(count or 0)
+    if item_count == 0:
+        return 0, None
+    return item_count, _coerce_datetime(newest)
+
+
+async def upsert_earth_news_items(db: AsyncSession, items: list[ParsedNewsItem]) -> int:
+    if not items:
+        return 0
+    now = datetime.now(UTC)
+    existing_result = await db.execute(
+        select(EarthNewsItem).where(EarthNewsItem.id.in_([item.id for item in items]))
+    )
+    existing = {record.id: record for record in existing_result.scalars().all()}
+    changed = 0
+    for item in items:
+        record = existing.get(item.id)
+        if record is None:
+            patch = build_anchor_location_patch(item)
+            record = EarthNewsItem(
+                id=item.id,
+                title=item.title,
+                summary=item.summary,
+                content_language=item.content_language,
+                localizations=dict(item.localizations or {}),
+                url=item.url,
+                source=item.source,
+                feed_name=item.feed_name,
+                region=item.feed_region,
+                homepage_url=item.homepage_url,
+                published_at=item.published_at,
+                latitude=patch["latitude"],
+                longitude=patch["longitude"],
+                location_label=patch["location_label"],
+                location_source=patch["location_source"],
+                verified=patch["verified"],
+                location_meta=patch["location_meta"],
+                first_seen_at=now,
+                last_seen_at=now,
+                enrichment_status=item.enrichment_status,
+                enrichment_error=item.enrichment_error,
+                enriched_at=item.enriched_at,
+            )
+            db.add(record)
+            changed += 1
+            continue
+
+        record.title = item.title
+        record.summary = item.summary
+        record.url = item.url
+        record.source = item.source
+        record.feed_name = item.feed_name
+        record.region = item.feed_region
+        record.homepage_url = item.homepage_url
+        record.published_at = item.published_at
+        record.last_seen_at = now
+        if item.localizations:
+            record.content_language = item.content_language
+            record.localizations = dict(item.localizations or {})
+            record.enrichment_status = item.enrichment_status
+            record.enrichment_error = item.enrichment_error
+            record.enriched_at = item.enriched_at
+        changed += 1
+    await db.flush()
+    return changed
+
+
+async def update_earth_news_item_location(
+    db: AsyncSession,
+    *,
+    item_id: str,
+    patch: dict[str, Any],
+) -> bool:
+    record = await db.get(EarthNewsItem, item_id)
+    if record is None:
+        return False
+    record.latitude = float(patch["latitude"])
+    record.longitude = float(patch["longitude"])
+    record.location_label = str(patch["location_label"])
+    record.location_source = str(patch["location_source"])
+    record.verified = bool(patch["verified"])
+    record.location_meta = dict(patch.get("location_meta") or {})
+    record.resolved_at = datetime.now(UTC) if record.verified else None
+    await db.flush()
+    return True
+
+
+async def update_earth_news_item_enrichment(
+    db: AsyncSession,
+    *,
+    item_id: str,
+    patch: dict[str, Any],
+) -> bool:
+    record = await db.get(EarthNewsItem, item_id)
+    if record is None:
+        return False
+    if "latitude" in patch:
+        record.latitude = float(patch["latitude"])
+        record.longitude = float(patch["longitude"])
+        record.location_label = str(patch["location_label"])
+        record.location_source = str(patch["location_source"])
+        record.verified = bool(patch["verified"])
+        record.location_meta = dict(patch.get("location_meta") or {})
+        record.resolved_at = datetime.now(UTC) if record.verified else None
+    if "content_language" in patch:
+        record.content_language = str(patch.get("content_language") or "en")
+    if "localizations" in patch:
+        record.localizations = dict(patch.get("localizations") or {})
+    if "enrichment_status" in patch:
+        record.enrichment_status = str(patch.get("enrichment_status") or "pending")
+    if "enrichment_error" in patch:
+        record.enrichment_error = patch.get("enrichment_error")
+    if patch.get("enriched_at"):
+        try:
+            parsed_enriched_at = datetime.fromisoformat(
+                str(patch["enriched_at"]).replace("Z", "+00:00")
+            )
+        except ValueError:
+            parsed_enriched_at = datetime.now(UTC)
+        record.enriched_at = _coerce_datetime(parsed_enriched_at)
+    elif patch.get("localizations"):
+        record.enriched_at = datetime.now(UTC)
+    await db.flush()
+    return True
+
+
+async def list_unverified_earth_news_items(
+    db: AsyncSession,
+    *,
+    active_region: str,
+    limit: int,
+) -> list[ParsedNewsItem]:
+    regions = {"global", active_region}
+    result = await db.execute(
+        select(EarthNewsItem)
+        .where(EarthNewsItem.region.in_(regions))
+        .where(EarthNewsItem.verified.is_(False))
+        .order_by(*_query_sort_key(active_region))
+        .limit(limit)
+    )
+    return [record_to_parsed_news_item(record) for record in result.scalars().all()]
+
+
+async def list_all_earth_news_records(db: AsyncSession) -> list[EarthNewsItem]:
+    result = await db.execute(
+        select(EarthNewsItem).order_by(
+            EarthNewsItem.published_at.desc().nullslast(),
+            EarthNewsItem.last_seen_at.desc(),
+        )
+    )
+    return list(result.scalars().all())
--- a/backend/app/services/earth_news_worker.py
+++ b/backend/app/services/earth_news_worker.py
@@ -0,0 +1,142 @@
+from __future__ import annotations
+
+import asyncio
+from contextlib import suppress
+from socket import gethostname
+from typing import Any
+
+from app.core.logging import get_logger
+from app.core.websocket.broadcaster import broadcaster
+from app.db.session import async_session_factory
+from app.services.ai_client import AIProviderClient
+from app.ai_tasks.prompts import get_effective_prompt
+from app.services.earth_news import (
+    NEWS_ENRICH_PROMPT_KEY,
+    _infer_news_enrichment,
+    build_target_location_patch,
+    parsed_news_item_from_job_payload,
+)
+from app.services.earth_news_queue import (
+    NewsTargetLocationMessage,
+    get_news_target_location_queue,
+    save_target_location_patch,
+)
+from app.services.earth_news_store import update_earth_news_item_enrichment as update_earth_news_item_location
+
+
+logger = get_logger(__name__, service="earth_news")
+
+WORKER_BATCH_SIZE = 4
+WORKER_BLOCK_MS = 5000
+WORKER_BACKOFF_SECONDS = 5.0
+
+_worker_task: asyncio.Task | None = None
+
+
+async def _build_provider_client() -> AIProviderClient | None:
+    try:
+        from app.api.v1.settings import get_runtime_ai_provider_config
+
+        async with async_session_factory() as session:
+            runtime_config = await get_runtime_ai_provider_config(session)
+        return AIProviderClient(
+            service_url=runtime_config["service_url"],
+            service_token=runtime_config["service_token"],
+            timeout=runtime_config["timeout_seconds"],
+            retry_attempts=runtime_config["retry_attempts"],
+            llm_config=runtime_config.get("llm_config") or {},
+        )
+    except Exception as exc:
+        logger.warning_event(
+            "Failed to build Earth news AI provider client",
+            event="earth_news.target_location.provider_unavailable",
+            context={"error": str(exc)},
+        )
+        return None
+
+
+async def process_target_location_message(
+    message: NewsTargetLocationMessage,
+    *,
+    provider_client: AIProviderClient | None,
+) -> dict[str, Any]:
+    item = parsed_news_item_from_job_payload(message.payload)
+    async with async_session_factory() as session:
+        prompt = await get_effective_prompt(session, NEWS_ENRICH_PROMPT_KEY)
+    target, localizations = await _infer_news_enrichment(
+        item,
+        provider_client=provider_client,
+        prompt=prompt,
+    )
+    item.target_location = target
+    item.localizations = localizations or item.localizations
+    patch = build_target_location_patch(item, target)
+    await save_target_location_patch(item.id, patch)
+    async with async_session_factory() as session:
+        await update_earth_news_item_location(session, item_id=item.id, patch=patch)
+        await session.commit()
+    await broadcaster.broadcast_custom(
+        "earth_news",
+        {
+            "item_id": item.id,
+            "patch": patch,
+        },
+    )
+    return patch
+
+
+async def _run_target_location_worker() -> None:
+    consumer_name = f"{gethostname()}:{id(asyncio.current_task())}"
+    queue = get_news_target_location_queue()
+    while True:
+        try:
+            messages = await queue.consume_batch(
+                consumer_name=consumer_name,
+                count=WORKER_BATCH_SIZE,
+                block_ms=WORKER_BLOCK_MS,
+            )
+        except asyncio.CancelledError:
+            raise
+        except Exception as exc:
+            logger.warning_event(
+                "Earth news target location worker queue read failed",
+                event="earth_news.target_location.worker_read_failed",
+                context={"error": str(exc)},
+            )
+            await asyncio.sleep(WORKER_BACKOFF_SECONDS)
+            continue
+
+        if not messages:
+            continue
+        provider_client = await _build_provider_client()
+        for message in messages:
+            try:
+                await process_target_location_message(message, provider_client=provider_client)
+                await queue.ack(message.message_id)
+            except asyncio.CancelledError:
+                raise
+            except Exception as exc:
+                logger.warning_event(
+                    "Earth news target location worker job failed",
+                    event="earth_news.target_location.worker_job_failed",
+                    context={"item_id": message.item_id, "error": str(exc)},
+                )
+                with suppress(Exception):
+                    await queue.retry_or_dead_letter(message, error=str(exc))
+
+
+def start_earth_news_target_worker() -> None:
+    global _worker_task
+    if _worker_task is None or _worker_task.done():
+        _worker_task = asyncio.create_task(_run_target_location_worker())
+
+
+async def stop_earth_news_target_worker() -> None:
+    global _worker_task
+    task = _worker_task
+    if task is None:
+        return
+    task.cancel()
+    with suppress(asyncio.CancelledError):
+        await task
+    _worker_task = None
--- a/backend/app/services/location/llm_fallback.py
+++ b/backend/app/services/location/llm_fallback.py
@@ -7,8 +7,11 @@ import re
 from dataclasses import dataclass
 from typing import Any, Iterable

+from sqlalchemy.ext.asyncio import AsyncSession
+
 from app.core.countries import COUNTRY_ENTRIES, normalize_country
 from app.schemas.ai import SituationalAnalysisRequest
+from app.ai_tasks.prompts import get_effective_prompt
 from app.services.ai_client import AIProviderClient
 from app.services.ai_tools.evidence_store import normalize_search_evidence
 from app.services.ai_tools.web_search import WebSearchClient, WebSearchError
@@ -23,6 +26,8 @@ from app.services.location.text import (

 VALID_LLM_PRECISIONS = {"precise", "site", "city"}
 DEFAULT_MIN_CONFIDENCE = 0.55
+LOCATION_NORMALIZE_PROMPT_KEY = "location.factcheck.normalize"
+LOCATION_RESOLVE_PROMPT_KEY = "location.factcheck.resolve"
 MODEL_CONFIDENCE_WEIGHT = 0.25
 _geocode_llm_city = build_default_nominatim_geocoder()
 _LLM_LOCATION_NAME_KEYS = (
@@ -876,6 +881,7 @@ async def _repair_location_payload_from_text(
    raw_text: str,
    query: LocationQuery,
    entity_type: str,
+    db: AsyncSession | None = None,
 ) -> dict[str, Any] | None:
    """Second-pass structure repair for models that answer in prose.

@@ -884,12 +890,11 @@ async def _repair_location_payload_from_text(
    """
    if not coerce_str(raw_text):
        return None
+    prompt = await get_effective_prompt(db, LOCATION_NORMALIZE_PROMPT_KEY)
    request = SituationalAnalysisRequest(
        title=f"Normalize location factcheck for {entity_type}",
-        objective=(
-            "Convert the supplied location factcheck text into exactly one strict "
-            "JSON object. Extract only facts present in the text or original query."
-        ),
+        objective=prompt.prompt,
+        system_prompt=prompt.system_prompt or None,
        context={
            "entity_type": entity_type,
            "location_query": _query_context(query),
@@ -929,6 +934,7 @@ async def collect_llm_location_fallback_candidate(
    provider_client: AIProviderClient,
    query: LocationQuery,
    entity_type: str,
+    db: AsyncSession | None = None,
    attempted_queries: Iterable[str] = (),
    search_evidence: list[dict[str, Any]] | None = None,
    min_confidence: float = DEFAULT_MIN_CONFIDENCE,
@@ -946,13 +952,11 @@ async def collect_llm_location_fallback_candidate(
            attempted_queries=[attempt],
            failure_reason="LLM location factcheck skipped: no WebSearch evidence.",
        )
+    prompt = await get_effective_prompt(db, LOCATION_RESOLVE_PROMPT_KEY)
    request = SituationalAnalysisRequest(
        title=f"Location factcheck fallback for {entity_type}",
-        objective=(
-            "Return exactly one JSON object for the most likely physical location. "
-            "Use only fact-checkable public knowledge; return null fields rather "
-            "than guessing when evidence is weak."
-        ),
+        objective=prompt.prompt,
+        system_prompt=prompt.system_prompt or None,
        context={
            "entity_type": entity_type,
            "location_query": _query_context(query),
@@ -1001,6 +1005,7 @@ async def collect_llm_location_fallback_candidate(
            raw_text=response.content,
            query=query,
            entity_type=entity_type,
+            db=db,
        )
    if payload is None:
        payload = _payload_from_free_text(response.content, query=query)
--- a/backend/app/services/scheduler.py
+++ b/backend/app/services/scheduler.py
@@ -175,19 +175,29 @@ async def run_collector_task(collector_name: str):
            )

        try:
-            collector._datasource_id = datasource.id
+            datasource_id = datasource.id
+            datasource_source = datasource.source
+            collector._datasource_id = datasource_id
            logger.info_event(
                "Running collector",
                event="collector.run.started",
-                context={"collector_name": collector_name, "datasource_id": datasource.id},
+                context={"collector_name": collector_name, "datasource_id": datasource_id},
            )
            task_result = await collector.run(db)
+            datasource = await db.get(DataSource, datasource_id)
+            if datasource is None:
+                logger.error_event(
+                    "Datasource disappeared after collector run",
+                    event="collector.run.datasource_missing_after_run",
+                    context={"collector_name": collector_name, "datasource_id": datasource_id},
+                )
+                return
            datasource.last_run_at = datetime.now(UTC)
            datasource.last_status = task_result.get("status")
            if datasource.last_status == "success":
-                effective_candidate = await get_builtin_effective_candidate(db, datasource.source)
+                effective_candidate = await get_builtin_effective_candidate(db, datasource_source)
                checksum, _credential_context = await build_builtin_connectivity_checksum(
-                    datasource.source,
+                    datasource_source,
                    effective_candidate["endpoint"],
                    effective_candidate["auth_type"],
                    effective_candidate["headers"],
@@ -196,7 +206,7 @@ async def run_collector_task(collector_name: str):
                )
                await save_connectivity_success(
                    db,
-                    datasource.source,
+                    datasource_source,
                    checksum,
                    {"status_code": None},
                    connected_by="collection",
@@ -205,9 +215,11 @@ async def run_collector_task(collector_name: str):
            logger.info_event(
                "Collector completed",
                event="collector.run.completed",
-                context={"collector_name": collector_name, "datasource_id": datasource.id, "result": task_result},
+                context={"collector_name": collector_name, "datasource_id": datasource_id, "result": task_result},
            )
        except asyncio.CancelledError:
+            await db.rollback()
+            datasource = await db.get(DataSource, datasource_id)
            datasource.last_run_at = datetime.now(UTC)
            datasource.last_status = "cancelled"
            await db.commit()
@@ -218,6 +230,8 @@ async def run_collector_task(collector_name: str):
            )
            raise
        except Exception as exc:
+            await db.rollback()
+            datasource = await db.get(DataSource, datasource_id)
            datasource.last_run_at = datetime.now(UTC)
            datasource.last_status = "failed"
            await db.commit()
@@ -374,6 +388,11 @@ def run_collector_now(collector_name: str) -> bool:
        return False


+def is_collector_running(collector_name: str) -> bool:
+    task = get_running_collector_task(collector_name)
+    return bool(task is not None and not task.done())
+
+
 async def cancel_running_collector_now(collector_name: str) -> bool:
    task = get_running_collector_task(collector_name)
    if task is None or task.done():
--- a/backend/app/services/situational_alert_ai_brief.py
+++ b/backend/app/services/situational_alert_ai_brief.py
@@ -10,8 +10,11 @@ from app.models.alert import Alert, AlertSeverity, AlertStatus
 from app.models.bgp_anomaly import BGPAnomaly
 from app.models.bgp_incident import BGPIncident
 from app.schemas.ai import SituationalAnalysisRequest
+from app.ai_tasks.prompts import get_effective_prompt
 from app.services.bgp_ai_brief_store import get_latest_bgp_brief_record

+SITUATIONAL_ALERT_BRIEF_PROMPT_KEY = "alerts.situational.brief"
+

 def _format_pairs(pairs: list[tuple[str, int]], empty_text: str = "无") -> str:
    if not pairs:
@@ -96,18 +99,24 @@ async def build_situational_alert_brief_request(
        (str(item[0] or "未命名数据源"), item[1])
        for item in alert_source_result.fetchall()
    ]
+    total_alerts = total_alerts_result.scalar() or 0
+    active_alerts = active_alerts_result.scalar() or 0
+    total_incidents = total_incidents_result.scalar() or 0
+    active_incidents = active_incidents_result.scalar() or 0
+    total_anomalies = total_anomalies_result.scalar() or 0
+    active_anomalies = active_anomalies_result.scalar() or 0

    facts = [
        (
-            f"系统告警侧：总告警 {total_alerts_result.scalar() or 0} 条，active {active_alerts_result.scalar() or 0} 条；"
+            f"系统告警侧：总告警 {total_alerts} 条，active {active_alerts} 条；"
            f"活跃告警严重度分布为 {_format_pairs(active_alert_severities)}。"
        ),
        (
-            f"BGP态势侧：累计 incidents {total_incidents_result.scalar() or 0} 条，active incidents {active_incidents_result.scalar() or 0} 条；"
+            f"BGP态势侧：累计 incidents {total_incidents} 条，active incidents {active_incidents} 条；"
            f"活跃 incidents 严重度分布为 {_format_pairs(active_bgp_severities)}。"
        ),
        (
-            f"BGP异常侧：累计 anomalies {total_anomalies_result.scalar() or 0} 条，active anomalies {active_anomalies_result.scalar() or 0} 条；"
+            f"BGP异常侧：累计 anomalies {total_anomalies} 条，active anomalies {active_anomalies} 条；"
            f"活跃 anomaly 类型分布为 {_format_pairs(active_anomaly_types)}。"
        ),
    ]
@@ -147,21 +156,23 @@ async def build_situational_alert_brief_request(

    context = {
        "source": "situational-alerts",
-        "active_system_alerts": active_alerts_result.scalar() or 0,
+        "active_system_alerts": active_alerts,
        "active_system_alert_severities": dict(active_alert_severities),
        "top_system_alert_sources": dict(active_alert_sources),
-        "active_bgp_incidents": active_incidents_result.scalar() or 0,
+        "active_bgp_incidents": active_incidents,
        "active_bgp_incident_severities": dict(active_bgp_severities),
-        "active_bgp_anomalies": active_anomalies_result.scalar() or 0,
+        "active_bgp_anomalies": active_anomalies,
        "active_bgp_anomaly_types": dict(active_anomaly_types),
        "bgp_hot_regions": dict(bgp_region_counter.most_common(5)),
        "latest_bgp_brief_id": latest_bgp_brief.id if latest_bgp_brief else None,
        "latest_bgp_brief_generated_at": latest_bgp_brief.generated_at if latest_bgp_brief else None,
    }
+    prompt = await get_effective_prompt(db, SITUATIONAL_ALERT_BRIEF_PROMPT_KEY)

    request = SituationalAnalysisRequest(
        title="态势告警 AI 简报",
-        objective="综合系统告警、BGP incidents、BGP anomalies 与近期 BGP AI 简报，生成一份面向值班人员的态势告警简报，指出当前最需要关注的风险域、跨模块联动迹象和优先动作。",
+        objective=prompt.prompt,
+        system_prompt=prompt.system_prompt or None,
        observations=facts,
        constraints=[
            "明确区分事实、推断与建议。",
--- a/backend/tests/test_api.py
+++ b/backend/tests/test_api.py
@@ -5,10 +5,12 @@ from datetime import datetime
 from unittest.mock import patch, AsyncMock
 from httpx import AsyncClient, ASGITransport

+from app.api.v1 import earth as earth_api
 from app.main import app
 from app.core.config import settings
 from app.core.security import create_access_token
 from app.db.session import get_db
+from app.models.system_setting import SystemSetting
 from app.models.user import User
 from app.schemas.ai import (
    AIProviderStatusResponse,
@@ -18,6 +20,17 @@ from app.schemas.ai import (
 )


+class _FakeRedisClient:
+    def sismember(self, *_args, **_kwargs):
+        return False
+
+
+@pytest.fixture(autouse=True)
+def fake_token_blacklist(monkeypatch):
+    """Keep API auth tests independent from an external Redis service."""
+    monkeypatch.setattr("app.core.security.redis_client", _FakeRedisClient())
+
+
@pytest.fixture
 def auth_headers():
    """Create authentication headers"""
@@ -50,6 +63,148 @@ async def test_root_endpoint():
    assert data["version"] == settings.VERSION


+class _ScalarOneOrNoneResult:
+    def __init__(self, value=None):
+        self._value = value
+
+    def scalar_one_or_none(self):
+        return self._value
+
+
+class _FakeEarthBrandSession:
+    def __init__(self, record=None):
+        self.record = record
+        self.added = None
+        self.deleted = False
+        self.committed = False
+
+    async def execute(self, statement):
+        if statement.__class__.__name__ == "Delete":
+            self.deleted = True
+            self.record = None
+            return _ScalarOneOrNoneResult(None)
+        return _ScalarOneOrNoneResult(self.record)
+
+    def add(self, record):
+        self.added = record
+        self.record = record
+
+    async def commit(self):
+        self.committed = True
+
+    async def refresh(self, _record):
+        return None
+
+
+def _override_admin_user():
+    return User(id=1, username="testuser", email="test@example.com", role="admin", is_active=True)
+
+
+@pytest.mark.asyncio
+async def test_get_earth_brand_returns_static_defaults():
+    async def override_get_db():
+        yield _FakeEarthBrandSession()
+
+    app.dependency_overrides[get_db] = override_get_db
+    try:
+        transport = ASGITransport(app=app)
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.get("/api/v1/earth/brand")
+        assert response.status_code == 200
+        data = response.json()
+        assert data["is_default"] is True
+        assert data["brand"]["logo_src"] == "/earth/assets/brand/earth-logo.png"
+        assert data["brand"]["title_src"] == "/earth/assets/brand/title-zh.png"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_update_and_reset_earth_brand(auth_headers):
+    session = _FakeEarthBrandSession()
+
+    async def override_get_db():
+        yield session
+
+    app.dependency_overrides.update(
+        {
+            __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: _override_admin_user,
+            get_db: override_get_db,
+        }
+    )
+    try:
+        transport = ASGITransport(app=app)
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            update_response = await client.put(
+                "/api/v1/earth/brand",
+                headers=auth_headers,
+                json={
+                    "logo_src": "/earth-brand-assets/custom.png",
+                    "title_src": "",
+                    "title_text": "Custom Earth",
+                    "subtitle": "Custom subtitle",
+                    "description": "Custom description",
+                    "aria_label": "",
+                    "title_alt": "",
+                },
+            )
+            reset_response = await client.delete("/api/v1/earth/brand", headers=auth_headers)
+
+        assert update_response.status_code == 200
+        updated = update_response.json()
+        assert updated["is_default"] is False
+        assert updated["brand"]["title_text"] == "Custom Earth"
+        assert updated["brand"]["aria_label"] == "Custom Earth"
+        assert isinstance(session.added, SystemSetting)
+        assert reset_response.status_code == 200
+        reset = reset_response.json()
+        assert reset["is_default"] is True
+        assert reset["brand"]["logo_src"] == "/earth/assets/brand/earth-logo.png"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_upload_earth_brand_asset_rejects_invalid_type(auth_headers):
+    app.dependency_overrides[
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user
+    ] = _override_admin_user
+    try:
+        transport = ASGITransport(app=app)
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.post(
+                "/api/v1/earth/brand/assets",
+                headers=auth_headers,
+                files={"file": ("brand.txt", b"nope", "text/plain")},
+            )
+        assert response.status_code == 400
+        assert response.json()["detail"]["code"] == "unsupported_file_type"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_upload_earth_brand_asset_saves_file(auth_headers, tmp_path, monkeypatch):
+    monkeypatch.setattr(earth_api, "EARTH_BRAND_ASSET_DIR", tmp_path)
+    app.dependency_overrides[
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user
+    ] = _override_admin_user
+    try:
+        transport = ASGITransport(app=app)
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.post(
+                "/api/v1/earth/brand/assets",
+                headers=auth_headers,
+                files={"file": ("brand.png", b"png-bytes", "image/png")},
+            )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["url"].startswith("/earth-brand-assets/")
+        assert (tmp_path / data["filename"]).read_bytes() == b"png-bytes"
+    finally:
+        app.dependency_overrides.clear()
+
+
@pytest.mark.asyncio
 async def test_dashboard_stats_without_auth():
    """Test dashboard stats requires authentication"""
@@ -62,20 +217,43 @@ async def test_dashboard_stats_without_auth():
@pytest.mark.asyncio
 async def test_dashboard_stats_with_auth(auth_headers):
    """Test dashboard stats with authentication"""
-    with patch("app.api.v1.dashboard.cache.get", return_value=None):
-        with patch("app.api.v1.dashboard.cache.set", return_value=True):
-            with patch("app.db.session.get_db") as mock_get_db:
-                mock_session = AsyncMock()
-                mock_result = AsyncMock()
-                mock_result.scalar.return_value = 0
-                mock_result.fetchall.return_value = []
-                mock_session.execute.return_value = mock_result
+    class _StatsResult:
+        def __init__(self, row):
+            self._row = row

-                async def mock_db_context():
-                    yield mock_session
+        def one(self):
+            return self._row

-                mock_get_db.return_value = mock_db_context()
+    class _FakeStatsSession:
+        def __init__(self):
+            self._rows = [
+                type("DatasourceStats", (), {"custom_count": 0, "custom_active": 0})(),
+                type("TaskStats", (), {"tasks_today": 0, "success_tasks": 0})(),
+                type(
+                    "AlertStats",
+                    (),
+                    {"critical_alerts": 0, "warning_alerts": 0, "info_alerts": 0},
+                )(),
+            ]

+        async def execute(self, _query):
+            return _StatsResult(self._rows.pop(0))
+
+    def override_get_current_user():
+        return User(id=1, username="testuser", email="test@example.com", role="admin", is_active=True)
+
+    async def override_get_db():
+        yield _FakeStatsSession()
+
+    app.dependency_overrides.update(
+        {
+            __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+            get_db: override_get_db,
+        }
+    )
+    try:
+        with patch("app.api.v1.dashboard.cache.get", return_value=None):
+            with patch("app.api.v1.dashboard.cache.set", return_value=True):
                transport = ASGITransport(app=app)
                async with AsyncClient(transport=transport, base_url="http://test") as client:
                    response = await client.get(
@@ -85,6 +263,8 @@ async def test_dashboard_stats_with_auth(auth_headers):
                assert response.status_code == 200
                data = response.json()
                assert "total_datasources" in data
+    finally:
+        app.dependency_overrides.clear()


@pytest.mark.asyncio
--- a/backend/tests/test_earth_boundaries.py
+++ b/backend/tests/test_earth_boundaries.py
@@ -0,0 +1,177 @@
+import json
+
+import pytest
+
+from app.core.datasource_defaults import DEFAULT_DATASOURCES
+from app.services.collectors.registry import collector_registry
+from app.services import earth_boundaries
+
+
+def write_geojson(path, name="Test"):
+    path.write_text(
+        json.dumps(
+            {
+                "type": "FeatureCollection",
+                "features": [
+                    {
+                        "type": "Feature",
+                        "properties": {"name": name},
+                        "geometry": {
+                            "type": "Polygon",
+                            "coordinates": [[[-1, -1], [1, -1], [1, 1], [-1, 1], [-1, -1]]],
+                        },
+                    }
+                ],
+            }
+        ),
+        encoding="utf-8",
+    )
+
+
+def patch_paths(monkeypatch, tmp_path):
+    repo = tmp_path
+    source_dir = repo / "data/earth-boundary-sources"
+    boundary_dir = repo / "frontend/public/earth/data/boundaries/v1"
+    pmtiles = repo / "frontend/public/earth/data/boundaries/earth-boundaries-china-pov-v1.pmtiles"
+    legacy = repo / "frontend/public/earth/data/countries-admin0.min.geojson"
+    config = repo / "config/earth-boundary-sources.local.json"
+    example = repo / "config/earth-boundary-sources.example.json"
+    policy = repo / "config/earth-boundary-pov-policy.china-v1.json"
+    for path in (source_dir, boundary_dir, pmtiles.parent, legacy.parent, config.parent):
+        path.mkdir(parents=True, exist_ok=True)
+    policy.write_text('{"productionTileFormat":"pmtiles+mvt"}\n', encoding="utf-8")
+    example.write_text('{"collectorConfigs":{}}\n', encoding="utf-8")
+    monkeypatch.setattr(earth_boundaries, "REPO_ROOT", repo)
+    monkeypatch.setattr(earth_boundaries, "SOURCE_OUTPUT_DIR", source_dir)
+    monkeypatch.setattr(earth_boundaries, "SOURCE_MANIFEST_PATH", source_dir / "manifest.json")
+    monkeypatch.setattr(earth_boundaries, "BUILD_RESULT_PATH", source_dir / "build-result.json")
+    monkeypatch.setattr(earth_boundaries, "BUILD_JOB_PATH", source_dir / "build-job.json")
+    monkeypatch.setattr(earth_boundaries, "BOUNDARY_OUTPUT_DIR", boundary_dir)
+    monkeypatch.setattr(earth_boundaries, "BOUNDARY_MANIFEST_PATH", boundary_dir / "manifest.json")
+    monkeypatch.setattr(earth_boundaries, "PMTILES_ARTIFACT_PATH", pmtiles)
+    monkeypatch.setattr(earth_boundaries, "LEGACY_GEOJSON_PATH", legacy)
+    monkeypatch.setattr(earth_boundaries, "LOCAL_CONFIG_PATH", config)
+    monkeypatch.setattr(earth_boundaries, "EXAMPLE_CONFIG_PATH", example)
+    monkeypatch.setattr(earth_boundaries, "POV_POLICY_PATH", policy)
+    return {
+        "repo": repo,
+        "config": config,
+        "legacy": legacy,
+        "pmtiles": pmtiles,
+        "manifest": boundary_dir / "manifest.json",
+    }
+
+
+def test_boundary_status_uses_legacy_provider_when_pmtiles_missing(monkeypatch, tmp_path):
+    paths = patch_paths(monkeypatch, tmp_path)
+    write_geojson(paths["legacy"])
+
+    status = earth_boundaries.get_boundary_status()
+
+    assert status["provider"] == "legacy-geojson"
+    assert status["fallback_available"] is True
+    assert status["high_precision_ready"] is False
+
+
+def test_boundary_status_prefers_high_precision_when_manifest_and_pmtiles_exist(monkeypatch, tmp_path):
+    paths = patch_paths(monkeypatch, tmp_path)
+    write_geojson(paths["legacy"])
+    paths["pmtiles"].write_bytes(b"pmtiles")
+    paths["manifest"].write_text('{"tileProvider":"pmtiles-mvt"}\n', encoding="utf-8")
+
+    status = earth_boundaries.get_boundary_status()
+
+    assert status["provider"] == "pmtiles-mvt"
+    assert status["high_precision_ready"] is True
+
+
+def test_save_boundary_config_writes_local_config(monkeypatch, tmp_path):
+    paths = patch_paths(monkeypatch, tmp_path)
+    payload = {"collectorConfigs": {"earth_admin0_boundaries": {"endpoint": "file:///tmp/a.geojson"}}}
+
+    status = earth_boundaries.save_boundary_config(payload)
+
+    assert paths["config"].exists()
+    assert status["config_source"] == "local"
+    assert status["config"] == payload
+
+
+@pytest.mark.asyncio
+async def test_build_reports_missing_tools_after_source_artifacts(monkeypatch, tmp_path):
+    paths = patch_paths(monkeypatch, tmp_path)
+    source_files = {}
+    for source in earth_boundaries.BOUNDARY_SOURCE_KINDS:
+        source_path = paths["repo"] / f"{source}.geojson"
+        write_geojson(source_path, name=source)
+        source_files[source] = source_path
+    paths["config"].write_text(
+        json.dumps(
+            {
+                "collectorConfigs": {
+                    source: {
+                        "sourceKind": kind,
+                        "endpoint": str(source_files[source]),
+                        "method": "GET",
+                    }
+                    for source, kind in earth_boundaries.BOUNDARY_SOURCE_KINDS.items()
+                }
+            }
+        ),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(earth_boundaries.shutil, "which", lambda _tool: None)
+    progress_events = []
+
+    status = await earth_boundaries.build_boundary_assets(
+        lambda progress, phase, message, **_extra: progress_events.append((progress, phase, message))
+    )
+
+    assert status["provider"] == "geojson-high-precision"
+    assert status["high_precision_ready"] is True
+    assert (paths["repo"] / "data/earth-boundary-sources/manifest.json").exists()
+    assert paths["manifest"].exists()
+    assert any(phase == "download" for _progress, phase, _message in progress_events)
+
+
+@pytest.mark.asyncio
+async def test_start_boundary_build_job_records_geojson_fallback_success(monkeypatch, tmp_path):
+    paths = patch_paths(monkeypatch, tmp_path)
+    monkeypatch.setattr(earth_boundaries, "_build_task", None)
+    monkeypatch.setattr(earth_boundaries, "_build_job_state", {})
+    source_files = {}
+    for source in earth_boundaries.BOUNDARY_SOURCE_KINDS:
+        source_path = paths["repo"] / f"{source}.geojson"
+        write_geojson(source_path, name=source)
+        source_files[source] = source_path
+    paths["config"].write_text(
+        json.dumps(
+            {
+                "collectorConfigs": {
+                    source: {
+                        "sourceKind": kind,
+                        "endpoint": str(source_files[source]),
+                        "method": "GET",
+                    }
+                    for source, kind in earth_boundaries.BOUNDARY_SOURCE_KINDS.items()
+                }
+            }
+        ),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(earth_boundaries.shutil, "which", lambda _tool: None)
+
+    response = await earth_boundaries.start_boundary_build_job()
+    await earth_boundaries._build_task
+    status = earth_boundaries.get_boundary_build_status()
+
+    assert response["accepted"] is True
+    assert status["job"]["status"] == "succeeded"
+    assert status["job"]["result"]["provider"] == "geojson-high-precision"
+
+
+def test_earth_boundary_collectors_are_not_registered_as_datasources():
+    removed = set(earth_boundaries.BOUNDARY_SOURCE_KINDS) | {"earth_boundary_tiles"}
+
+    assert removed.isdisjoint(DEFAULT_DATASOURCES)
+    for source in removed:
+        assert collector_registry.get(source) is None
--- a/backend/tests/test_earth_news.py
+++ b/backend/tests/test_earth_news.py
@@ -1,6 +1,20 @@
 from datetime import UTC, datetime
+from types import SimpleNamespace

-from app.services.earth_news import ParsedNewsItem, _serialize_item
+import pytest
+
+from app.services.earth_news import (
+    NewsFeedSource,
+    NewsTargetLocation,
+    ParsedNewsItem,
+    _enrich_items_with_target_locations,
+    _extract_target_location_from_text,
+    _serialize_item,
+    get_earth_news_payload,
+)
+from app.services.earth_news_queue import NewsTargetLocationMessage
+from app.services.earth_news_worker import process_target_location_message
+from app.services.collectors.media_news_archive import MediaNewsArchiveCollector


 def test_serialize_item_includes_region_anchor_for_cruise():
@@ -21,7 +35,10 @@ def test_serialize_item_includes_region_anchor_for_cruise():
    assert payload["latitude"] == 1.3521
    assert payload["longitude"] == 103.8198
    assert payload["location_label"] == "亚太"
-    assert payload["location_inferred"] is True
+    assert payload["location_source"] == "region_anchor"
+    assert payload["verified"] is False
+    assert payload["location_meta"]["target"] is None
+    assert payload["location_meta"]["anchor"]["region"] == "asia-pacific"
    assert payload["is_focus_match"] is True
    assert payload["published_at"] == "2026-04-23T02:30:00Z"

@@ -44,6 +61,687 @@ def test_serialize_item_falls_back_to_global_anchor():
    assert payload["latitude"] == 20.0
    assert payload["longitude"] == 0.0
    assert payload["location_label"] == "全球"
-    assert payload["location_inferred"] is True
+    assert payload["location_source"] == "region_anchor"
+    assert payload["verified"] is False
+    assert payload["location_meta"]["anchor"]["region"] == "global"
    assert payload["is_focus_match"] is False
    assert payload["published_at"] is None
+
+
+def test_serialize_item_includes_inferred_target_location():
+    item = ParsedNewsItem(
+        id="bbc-world:f55310fb667b",
+        title="Watch: What happened on day one of Trump's China visit?",
+        summary=(
+            "China welcomed US President Donald Trump with cheering children "
+            "and a troop parade."
+        ),
+        url="https://example.com/china-visit",
+        source="BBC World",
+        feed_name="BBC World",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 7, 13, 24, tzinfo=UTC),
+        target_location=NewsTargetLocation(
+            latitude=39.9042,
+            longitude=116.4074,
+            label="Beijing, China",
+            source="ai_inferred_target",
+            confidence=0.88,
+            country="中国",
+            city="Beijing",
+        ),
+    )
+
+    payload = _serialize_item(item, active_region="global")
+
+    assert payload["latitude"] == 39.9042
+    assert payload["longitude"] == 116.4074
+    assert payload["location_label"] == "Beijing, China"
+    assert payload["location_source"] == "ai_inferred_target"
+    assert payload["verified"] is True
+    assert payload["location_meta"]["target"]["confidence"] == 0.88
+    assert payload["location_meta"]["target"]["country"] == "中国"
+    assert payload["location_meta"]["target"]["city"] == "Beijing"
+    assert payload["location_meta"]["resolution_stage"] == "unresolved"
+    assert payload["location_meta"]["ai_attempted"] is False
+    assert payload["location_meta"]["ai_status"] == "not_attempted"
+    assert payload["location_meta"]["ai_error"] is None
+
+
+@pytest.mark.asyncio
+async def test_enrich_items_with_target_locations_uses_ai_and_geocode(monkeypatch):
+    item = ParsedNewsItem(
+        id="bbc-world:f55310fb667b",
+        title="Watch: What happened on day one of Trump's China visit?",
+        summary="China welcomed US President Donald Trump before a long meeting with Xi Jinping.",
+        url="https://example.com/china-visit",
+        source="BBC World",
+        feed_name="BBC World",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 7, 13, 24, tzinfo=UTC),
+    )
+
+    async def fake_geocode(_query: str):
+        return {
+            "lat": "39.9042",
+            "lon": "116.4074",
+            "display_name": "Beijing, China",
+        }
+
+    class FakeProviderClient:
+        async def analyze(self, _request):
+            class Response:
+                content = (
+                    '{"country":"China","city":"Beijing","matched_location_name":"Beijing, China",'
+                    '"latitude":null,"longitude":null,"confidence":0.88}'
+                )
+
+            return Response()
+
+    monkeypatch.setattr("app.services.earth_news._geocode_target_location", fake_geocode)
+
+    enriched = await _enrich_items_with_target_locations(
+        [item],
+        provider_client=FakeProviderClient(),
+    )
+
+    assert len(enriched) == 1
+    assert enriched[0].target_location is not None
+    assert enriched[0].target_location.latitude == 39.9042
+    assert enriched[0].target_location.longitude == 116.4074
+    assert enriched[0].target_location.label == "Beijing, China"
+    assert enriched[0].target_resolution_stage == "ai_inferred_target"
+    assert enriched[0].target_ai_attempted is True
+    assert enriched[0].target_ai_status == "success"
+    assert enriched[0].target_ai_error is None
+
+
+@pytest.mark.asyncio
+async def test_enrich_items_with_target_locations_adds_localizations(monkeypatch):
+    item = ParsedNewsItem(
+        id="global-scan:localized",
+        title="Global leaders meet to discuss energy security",
+        summary="Officials said the talks focused on supply chains and grid resilience.",
+        url="https://example.com/energy-security",
+        source="Example Source",
+        feed_name="Global Monitor / World",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 7, 13, 24, tzinfo=UTC),
+    )
+
+    async def fake_geocode(_query: str):
+        return {
+            "lat": "50.1109",
+            "lon": "8.6821",
+            "display_name": "Frankfurt am Main, Germany",
+        }
+
+    class FakeProviderClient:
+        async def analyze(self, _request):
+            class Response:
+                content = (
+                    '{"location":{"country":"Germany","city":"Frankfurt",'
+                    '"matched_location_name":"Frankfurt, Germany",'
+                    '"latitude":null,"longitude":null,"confidence":0.77},'
+                    '"localizations":{"zh-CN":{"title":"全球领导人讨论能源安全",'
+                    '"summary":"官员表示，会谈聚焦供应链和电网韧性。"}}}'
+                )
+
+            return Response()
+
+    monkeypatch.setattr("app.services.earth_news._geocode_target_location", fake_geocode)
+
+    enriched = await _enrich_items_with_target_locations(
+        [item],
+        provider_client=FakeProviderClient(),
+    )
+    payload = _serialize_item(enriched[0], active_region="global")
+
+    assert payload["title"] == "Global leaders meet to discuss energy security"
+    assert payload["summary"] == "Officials said the talks focused on supply chains and grid resilience."
+    assert payload["localizations"]["zh-CN"]["title"] == "全球领导人讨论能源安全"
+    assert payload["display_title"] == "全球领导人讨论能源安全"
+    assert payload["display_summary"] == "官员表示，会谈聚焦供应链和电网韧性。"
+    assert payload["enrichment_status"] == "success"
+
+
+@pytest.mark.asyncio
+async def test_extract_target_location_from_text_uses_country_hint(monkeypatch):
+    item = ParsedNewsItem(
+        id="bbc-world:country-hint",
+        title="Giant new dinosaur identified from fossils in Thailand",
+        summary="The nagatitan is the largest dinosaur found in South-East Asia.",
+        url="https://example.com/thailand-dinosaur",
+        source="BBC World",
+        feed_name="BBC World",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 17, 28, 56, tzinfo=UTC),
+    )
+
+    target = await _extract_target_location_from_text(item)
+
+    assert target is not None
+    assert target.country == "泰国"
+    assert target.latitude == 15.87
+    assert target.longitude == 100.9925
+    assert target.source == "headline_country_hint"
+
+
+@pytest.mark.asyncio
+async def test_enrich_items_with_target_locations_records_ai_provider_error():
+    item = ParsedNewsItem(
+        id="global-scan:no-hint",
+        title="The New Geopolitics of Power: Whoever Controls Electrons Wins the Decade",
+        summary="A broad analysis of industrial policy and energy systems.",
+        url="https://example.com/geopolitics-power",
+        source="Example Source",
+        feed_name="Global Monitor / World",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 17, 3, 10, tzinfo=UTC),
+    )
+
+    class FailingProviderClient:
+        async def analyze(self, _request):
+            raise RuntimeError("upstream ai timeout")
+
+    enriched = await _enrich_items_with_target_locations(
+        [item],
+        provider_client=FailingProviderClient(),
+    )
+
+    assert len(enriched) == 1
+    assert enriched[0].target_location is None
+    assert enriched[0].target_resolution_stage == "unresolved"
+    assert enriched[0].target_ai_attempted is True
+    assert enriched[0].target_ai_status == "provider_error"
+    assert enriched[0].target_ai_error == "upstream ai timeout"
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_returns_anchor_items_and_enqueues_location_job(monkeypatch):
+    source = NewsFeedSource(
+        id="test-feed",
+        name="Test Feed",
+        region="global",
+        homepage_url="https://example.com",
+        feed_url="https://example.com/rss.xml",
+    )
+    item = ParsedNewsItem(
+        id="test-feed:timeout",
+        title="Example story",
+        summary="Example summary",
+        url="https://example.com/story",
+        source="Test Feed",
+        feed_name="Test Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+    )
+
+    async def fake_fetch_source(_client, feed_source):
+        return feed_source, [item], None
+
+    async def fake_get_cached_target_location_patch(_item_id):
+        return None
+
+    enqueued_payloads = []
+
+    async def fake_enqueue_target_location_job(payload, **_kwargs):
+        enqueued_payloads.append(payload)
+        return True
+
+    monkeypatch.setattr("app.services.earth_news.get_sources_for_region", lambda _region: [source])
+    monkeypatch.setattr("app.services.earth_news._fetch_source", fake_fetch_source)
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.get_cached_target_location_patch",
+        fake_get_cached_target_location_patch,
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.enqueue_target_location_job",
+        fake_enqueue_target_location_job,
+    )
+
+    payload = await get_earth_news_payload(provider_client=None)
+
+    assert len(payload["items"]) == 1
+    assert payload["items"][0]["id"] == "test-feed:timeout"
+    assert payload["items"][0]["display_title"] == ""
+    assert payload["items"][0]["display_summary"] == ""
+    assert payload["items"][0]["latitude"] == 20.0
+    assert payload["items"][0]["longitude"] == 0.0
+    assert payload["items"][0]["location_source"] == "region_anchor"
+    assert payload["items"][0]["verified"] is False
+    assert payload["items"][0]["location_meta"]["ai_status"] == "queued"
+    assert enqueued_payloads[0]["id"] == "test-feed:timeout"
+    assert payload["errors"] == []
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_uses_fresh_database_items_without_rss(monkeypatch):
+    db = object()
+    item = ParsedNewsItem(
+        id="db:fresh",
+        title="Fresh database story",
+        summary="Stored summary",
+        url="https://example.com/fresh",
+        source="Stored Source",
+        feed_name="Stored Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+    )
+    item.location_patch = {
+        "latitude": 39.9057136,
+        "longitude": 116.3912972,
+        "location_label": "北京市, 中国",
+        "location_source": "headline_location_hint",
+        "verified": True,
+        "location_meta": {"target": {"city": "Beijing"}, "anchor": {"region": "global"}},
+    }
+
+    async def fake_get_earth_news_freshness(_db, *, active_region):
+        return 12, datetime.now(UTC)
+
+    async def fake_list_earth_news_items(_db, *, active_region, limit):
+        assert limit == 12
+        return [item]
+
+    async def fail_fetch(_sources):
+        raise AssertionError("fresh database items should not fetch RSS")
+
+    monkeypatch.setattr("app.services.earth_news_store.get_earth_news_freshness", fake_get_earth_news_freshness)
+    monkeypatch.setattr("app.services.earth_news_store.list_earth_news_items", fake_list_earth_news_items)
+    monkeypatch.setattr("app.services.earth_news._fetch_rss_items_for_sources", fail_fetch)
+
+    payload = await get_earth_news_payload(db=db)
+
+    assert payload["items"][0]["id"] == "db:fresh"
+    assert payload["items"][0]["verified"] is True
+    assert payload["items"][0]["latitude"] == 39.9057136
+    assert payload["stale"] is False
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_initializes_empty_database_from_rss(monkeypatch):
+    db = object()
+    source = NewsFeedSource(
+        id="test-feed",
+        name="Test Feed",
+        region="global",
+        homepage_url="https://example.com",
+        feed_url="https://example.com/rss.xml",
+    )
+    item = ParsedNewsItem(
+        id="test-feed:init",
+        title="Initial RSS story",
+        summary="Initial summary",
+        url="https://example.com/init",
+        source="Test Feed",
+        feed_name="Test Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+    )
+    item.location_patch = {
+        "latitude": 20.0,
+        "longitude": 0.0,
+        "location_label": "全球",
+        "location_source": "region_anchor",
+        "verified": False,
+        "location_meta": {"target": None, "anchor": {"region": "global"}},
+    }
+    upserted = []
+    enqueued = []
+
+    async def fake_get_earth_news_freshness(_db, *, active_region):
+        return 0, None
+
+    async def fake_fetch_rss_items_for_sources(_sources):
+        return [item], []
+
+    async def fake_upsert_earth_news_items(_db, items):
+        upserted.extend(items)
+        return len(items)
+
+    async def fake_list_earth_news_items(_db, *, active_region, limit):
+        return [item]
+
+    async def fake_enqueue_target_location_job(payload, **_kwargs):
+        enqueued.append(payload)
+        return True
+
+    monkeypatch.setattr("app.services.earth_news.get_sources_for_region", lambda _region: [source])
+    monkeypatch.setattr("app.services.earth_news_store.get_earth_news_freshness", fake_get_earth_news_freshness)
+    monkeypatch.setattr("app.services.earth_news._fetch_rss_items_for_sources", fake_fetch_rss_items_for_sources)
+    monkeypatch.setattr("app.services.earth_news_store.upsert_earth_news_items", fake_upsert_earth_news_items)
+    monkeypatch.setattr("app.services.earth_news_store.list_earth_news_items", fake_list_earth_news_items)
+    monkeypatch.setattr("app.services.earth_news_queue.enqueue_target_location_job", fake_enqueue_target_location_job)
+
+    payload = await get_earth_news_payload(db=db)
+
+    assert upserted[0].id == "test-feed:init"
+    assert payload["items"][0]["id"] == "test-feed:init"
+    assert payload["items"][0]["verified"] is False
+    assert enqueued[0]["id"] == "test-feed:init"
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_supplements_stale_database_items(monkeypatch):
+    db = object()
+    old_item = ParsedNewsItem(
+        id="db:old",
+        title="Old story",
+        summary="Old summary",
+        url="https://example.com/old",
+        source="Stored Source",
+        feed_name="Stored Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 14, 3, 0, tzinfo=UTC),
+    )
+    old_item.location_patch = {
+        "latitude": 20.0,
+        "longitude": 0.0,
+        "location_label": "全球",
+        "location_source": "region_anchor",
+        "verified": False,
+        "location_meta": {"target": None, "anchor": {"region": "global"}},
+    }
+    fetched = []
+
+    async def fake_get_earth_news_freshness(_db, *, active_region):
+        return 12, datetime(2026, 5, 14, 3, 0, tzinfo=UTC)
+
+    async def fake_fetch_rss_items_for_sources(_sources):
+        fetched.append(True)
+        return [old_item], []
+
+    async def fake_upsert_earth_news_items(_db, items):
+        return len(items)
+
+    async def fake_list_earth_news_items(_db, *, active_region, limit):
+        return [old_item]
+
+    async def fake_enqueue_target_location_job(_payload, **_kwargs):
+        return True
+
+    monkeypatch.setattr("app.services.earth_news_store.get_earth_news_freshness", fake_get_earth_news_freshness)
+    monkeypatch.setattr("app.services.earth_news._fetch_rss_items_for_sources", fake_fetch_rss_items_for_sources)
+    monkeypatch.setattr("app.services.earth_news_store.upsert_earth_news_items", fake_upsert_earth_news_items)
+    monkeypatch.setattr("app.services.earth_news_store.list_earth_news_items", fake_list_earth_news_items)
+    monkeypatch.setattr("app.services.earth_news_queue.enqueue_target_location_job", fake_enqueue_target_location_job)
+
+    payload = await get_earth_news_payload(db=db)
+
+    assert fetched == [True]
+    assert payload["items"][0]["id"] == "db:old"
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_merges_cached_location_patch(monkeypatch):
+    source = NewsFeedSource(
+        id="test-feed",
+        name="Test Feed",
+        region="global",
+        homepage_url="https://example.com",
+        feed_url="https://example.com/rss.xml",
+    )
+    item = ParsedNewsItem(
+        id="test-feed:cached",
+        title="Cached story",
+        summary="Cached summary",
+        url="https://example.com/cached",
+        source="Test Feed",
+        feed_name="Test Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+    )
+    cached_patch = {
+        "latitude": 39.9057136,
+        "longitude": 116.3912972,
+        "location_label": "北京市, 中国",
+        "location_source": "headline_location_hint",
+        "verified": True,
+        "location_meta": {
+            "resolution_stage": "headline_location_hint",
+            "ai_attempted": False,
+            "ai_status": "skipped_text_hint",
+            "ai_error": None,
+            "debug_note": "text hint matched 北京市, 中国",
+            "target": {"city": "Beijing"},
+            "anchor": {"region": "global"},
+        },
+    }
+
+    async def fake_fetch_source(_client, feed_source):
+        return feed_source, [item], None
+
+    async def fake_get_cached_target_location_patch(_item_id):
+        return cached_patch
+
+    enqueued = []
+
+    async def fake_enqueue_target_location_job(payload, **_kwargs):
+        enqueued.append(payload)
+        return True
+
+    monkeypatch.setattr("app.services.earth_news.get_sources_for_region", lambda _region: [source])
+    monkeypatch.setattr("app.services.earth_news._fetch_source", fake_fetch_source)
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.get_cached_target_location_patch",
+        fake_get_cached_target_location_patch,
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.enqueue_target_location_job",
+        fake_enqueue_target_location_job,
+    )
+
+    payload = await get_earth_news_payload(provider_client=None)
+
+    assert payload["items"][0]["latitude"] == 39.9057136
+    assert payload["items"][0]["longitude"] == 116.3912972
+    assert payload["items"][0]["verified"] is True
+    assert payload["items"][0]["location_source"] == "headline_location_hint"
+    assert enqueued[0]["id"] == "test-feed:cached"
+
+
+@pytest.mark.asyncio
+async def test_earth_news_payload_requeues_cached_failed_localization(monkeypatch):
+    source = NewsFeedSource(
+        id="test-feed",
+        name="Test Feed",
+        region="global",
+        homepage_url="https://example.com",
+        feed_url="https://example.com/rss.xml",
+    )
+    item = ParsedNewsItem(
+        id="test-feed:failed-localization",
+        title="Failed localization story",
+        summary="English source summary.",
+        url="https://example.com/failed-localization",
+        source="Test Feed",
+        feed_name="Test Feed",
+        feed_region="global",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+    )
+    cached_patch = {
+        "latitude": 20.0,
+        "longitude": 0.0,
+        "location_label": "全球",
+        "location_source": "region_anchor",
+        "verified": False,
+        "location_meta": {"target": None, "anchor": {"region": "global"}},
+        "content_language": "en",
+        "localizations": {},
+        "enrichment_status": "parse_error",
+        "enrichment_error": "AI response did not contain a parseable JSON object.",
+        "enriched_at": None,
+    }
+    enqueued = []
+
+    async def fake_fetch_source(_client, feed_source):
+        return feed_source, [item], None
+
+    async def fake_get_cached_target_location_patch(_item_id):
+        return cached_patch
+
+    async def fake_enqueue_target_location_job(payload, **_kwargs):
+        enqueued.append(payload)
+        return True
+
+    monkeypatch.setattr("app.services.earth_news.get_sources_for_region", lambda _region: [source])
+    monkeypatch.setattr("app.services.earth_news._fetch_source", fake_fetch_source)
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.get_cached_target_location_patch",
+        fake_get_cached_target_location_patch,
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_queue.enqueue_target_location_job",
+        fake_enqueue_target_location_job,
+    )
+
+    payload = await get_earth_news_payload(provider_client=None)
+
+    assert enqueued[0]["id"] == "test-feed:failed-localization"
+    assert payload["items"][0]["display_title"] == ""
+    assert payload["items"][0]["enrichment_status"] == "queued"
+
+
+@pytest.mark.asyncio
+async def test_worker_processes_target_location_message_and_returns_patch(monkeypatch):
+    message = NewsTargetLocationMessage(
+        message_id="1-0",
+        item_id="bbc-world:worker",
+        payload={
+            "id": "bbc-world:worker",
+            "title": "Ukraine rescuers pull dead from rubble of Kyiv flats",
+            "summary": "Massive Russian drone and missile attacks in Ukraine's capital.",
+            "url": "https://example.com/kyiv",
+            "source": "BBC World",
+            "feed_name": "BBC World",
+            "feed_region": "global",
+            "homepage_url": "https://www.bbc.com/news/world",
+            "published_at": "2026-05-14T13:16:32Z",
+        },
+    )
+
+    async def fake_geocode(_query: str):
+        return {
+            "lat": "50.4500336",
+            "lon": "30.5241361",
+            "display_name": "Київ, Україна",
+        }
+
+    saved = {}
+    broadcasted = {}
+
+    async def fake_save_target_location_patch(item_id, patch):
+        saved["item_id"] = item_id
+        saved["patch"] = patch
+
+    async def fake_update_earth_news_item_location(_session, *, item_id, patch):
+        saved["db_item_id"] = item_id
+        saved["db_patch"] = patch
+        return True
+
+    async def fake_broadcast_custom(channel, data):
+        broadcasted["channel"] = channel
+        broadcasted["data"] = data
+
+    class FakeSession:
+        async def commit(self):
+            saved["committed"] = True
+
+    class FakeSessionFactory:
+        async def __aenter__(self):
+            return FakeSession()
+
+        async def __aexit__(self, exc_type, exc, tb):
+            return False
+
+    monkeypatch.setattr("app.services.earth_news._geocode_target_location", fake_geocode)
+    monkeypatch.setattr(
+        "app.services.earth_news_worker.save_target_location_patch",
+        fake_save_target_location_patch,
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_worker.update_earth_news_item_location",
+        fake_update_earth_news_item_location,
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_worker.async_session_factory",
+        lambda: FakeSessionFactory(),
+    )
+    monkeypatch.setattr(
+        "app.services.earth_news_worker.broadcaster.broadcast_custom",
+        fake_broadcast_custom,
+    )
+
+    patch = await process_target_location_message(message, provider_client=None)
+
+    assert patch["latitude"] == 50.4500336
+    assert patch["longitude"] == 30.5241361
+    assert patch["location_source"] == "headline_location_hint"
+    assert patch["verified"] is True
+    assert saved["item_id"] == "bbc-world:worker"
+    assert saved["db_item_id"] == "bbc-world:worker"
+    assert saved["committed"] is True
+    assert broadcasted["channel"] == "earth_news"
+    assert broadcasted["data"]["item_id"] == "bbc-world:worker"
+
+
+@pytest.mark.asyncio
+async def test_media_news_archive_collector_maps_news_items(monkeypatch):
+    collector = MediaNewsArchiveCollector()
+    collector._db_session = object()
+    record = SimpleNamespace(
+        id="bbc-world:archive",
+        title="Archived news",
+        summary="Archived summary",
+        url="https://example.com/archive",
+        source="BBC World",
+        feed_name="BBC World",
+        region="global",
+        homepage_url="https://www.bbc.com/news/world",
+        published_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+        latitude=39.9057136,
+        longitude=116.3912972,
+        location_label="北京市, 中国",
+        location_source="headline_location_hint",
+        verified=True,
+        location_meta={"target": {"country": "中国", "city": "Beijing"}},
+        content_language="en",
+        localizations={"zh-CN": {"title": "归档新闻", "summary": "归档概要"}},
+        enrichment_status="success",
+        enrichment_error=None,
+        enriched_at=datetime(2026, 5, 15, 3, 6, tzinfo=UTC),
+        first_seen_at=datetime(2026, 5, 15, 2, 0, tzinfo=UTC),
+        last_seen_at=datetime(2026, 5, 15, 3, 0, tzinfo=UTC),
+        resolved_at=datetime(2026, 5, 15, 3, 5, tzinfo=UTC),
+    )
+
+    async def fake_list_all_earth_news_records(_db):
+        return [record]
+
+    monkeypatch.setattr(
+        "app.services.collectors.media_news_archive.list_all_earth_news_records",
+        fake_list_all_earth_news_records,
+    )
+
+    items = await collector.fetch()
+
+    assert items[0]["source_id"] == "bbc-world:archive"
+    assert collector.data_type == "news_item"
+    assert items[0]["country"] == "中国"
+    assert items[0]["city"] == "Beijing"
+    assert items[0]["latitude"] == 39.9057136
+    assert items[0]["metadata"]["verified"] is True
+    assert "localizations" not in items[0]["metadata"]
+    assert "enrichment_status" not in items[0]["metadata"]
--- a/backend/tests/test_layers.py
+++ b/backend/tests/test_layers.py
@@ -1,4 +1,5 @@
 from fastapi import HTTPException
+import pytest

 from app.api.v1 import layers

@@ -42,3 +43,28 @@ def test_layer_guard_filters_bbox_and_clamps_low_zoom_limit():
    assert result["diagnostics"]["limit"] == layers.LOW_ZOOM_FEATURE_LIMIT
    assert result["diagnostics"]["limit_clamped"] is True
    assert result["diagnostics"]["degraded"] is True
+
+
+@pytest.mark.asyncio
+async def test_vessel_layer_snapshot_passes_type_filter(monkeypatch):
+    captured = {}
+
+    async def fake_build_vessel_snapshot_response(db, **kwargs):
+        captured.update(kwargs)
+        return {"type": "FeatureCollection", "features": []}
+
+    monkeypatch.setattr(layers, "build_vessel_snapshot_response", fake_build_vessel_snapshot_response)
+
+    result = await layers.get_vessel_layer_snapshot(
+        bbox="10,59,11,60",
+        zoom=12,
+        limit=1000,
+        vessel_type="cargo",
+        since_minutes=30,
+        db=object(),
+    )
+
+    assert result["features"] == []
+    assert captured["bbox"] == (10.0, 59.0, 11.0, 60.0)
+    assert captured["type_filter"] == "cargo"
+    assert "vessel_type" not in captured
--- a/backend/tests/test_realtime_sources.py
+++ b/backend/tests/test_realtime_sources.py
@@ -0,0 +1,101 @@
+from datetime import UTC, datetime
+from unittest.mock import AsyncMock
+
+import pytest
+from fastapi import HTTPException
+
+from app.api.v1 import realtime_sources
+from app.models.datasource import DataSource
+from app.models.datasource_config import DataSourceConfig
+from app.models.vessel import AISSourceHealth
+
+
+@pytest.mark.asyncio
+async def test_serialize_builtin_aisstream_includes_health_config_and_stats(monkeypatch):
+    datasource = DataSource(
+        id=28,
+        name="AISStream Vessels",
+        source="aisstream_vessels",
+        module="L4",
+        priority="P1",
+        collector_class="aisstream_vessels",
+        is_active=True,
+    )
+    config = DataSourceConfig(
+        id=3,
+        name="aisstream_vessels",
+        source_type="websocket",
+        endpoint="wss://stream.aisstream.io/v0/stream",
+        auth_type="api_key",
+        auth_config={"api_key": "test-key"},
+        config={
+            "message_types": ["PositionReport"],
+            "bounding_boxes": [[[-10, 50], [35, 75]]],
+        },
+        is_active=True,
+    )
+    health = AISSourceHealth(
+        source="aisstream_vessels",
+        connection_state="connected",
+        last_seen_at=datetime(2026, 5, 13, 1, 0, tzinfo=UTC),
+    )
+
+    class _Session:
+        async def get(self, _model, key):
+            assert key == "aisstream_vessels"
+            return health
+
+    monkeypatch.setattr(
+        realtime_sources,
+        "_load_realtime_stats",
+        AsyncMock(
+            return_value={
+                "total_observations": 10,
+                "observations_24h": 4,
+                "observations_1h": 1,
+                "unique_mmsi_total": 8,
+                "unique_mmsi_24h": 3,
+                "latest_observed_at": "2026-05-13T01:00:00Z",
+                "latest_collected_at": "2026-05-13T01:00:01Z",
+            }
+        ),
+    )
+    monkeypatch.setattr(realtime_sources, "is_collector_running", lambda source: False)
+
+    payload = await realtime_sources._serialize_builtin_aisstream(_Session(), datasource, config)
+
+    assert payload["source"] == "aisstream_vessels"
+    assert payload["kind"] == "builtin"
+    assert payload["credential_configured"] is True
+    assert payload["message_types"] == ["PositionReport"]
+    assert payload["runtime"]["running"] is False
+    assert payload["health"]["connection_state"] == "connected"
+    assert payload["stats"]["total_observations"] == 10
+
+
+@pytest.mark.asyncio
+async def test_start_builtin_realtime_source_rejects_disabled(monkeypatch):
+    datasource = DataSource(
+        id=28,
+        name="AISStream Vessels",
+        source="aisstream_vessels",
+        module="L4",
+        priority="P1",
+        collector_class="aisstream_vessels",
+        is_active=False,
+    )
+    monkeypatch.setattr(
+        realtime_sources,
+        "_load_builtin_aisstream",
+        AsyncMock(return_value=(datasource, None)),
+    )
+
+    with pytest.raises(HTTPException) as excinfo:
+        await realtime_sources.start_realtime_source(
+            "aisstream_vessels",
+            current_user=object(),
+            db=object(),
+        )
+
+    assert excinfo.value.status_code == 400
+    assert "disabled" in excinfo.value.detail
--- a/backend/tests/test_settings_ai_prompts.py
+++ b/backend/tests/test_settings_ai_prompts.py
@@ -0,0 +1,92 @@
+from types import SimpleNamespace
+
+import pytest
+
+from app.ai_tasks.prompts import (
+    get_effective_prompt,
+    list_effective_prompts,
+    reset_prompt_override,
+    save_prompt_override,
+)
+
+
+class _ScalarResult:
+    def __init__(self, value):
+        self._value = value
+
+    def scalar_one_or_none(self):
+        return self._value
+
+
+class _PromptSettingsDB:
+    def __init__(self, payload=None):
+        self.record = SimpleNamespace(category="ai_prompts", payload=payload) if payload is not None else None
+        self.added = None
+        self.commits = 0
+
+    async def execute(self, _statement):
+        return _ScalarResult(self.record)
+
+    def add(self, record):
+        self.record = record
+        self.added = record
+
+    async def commit(self):
+        self.commits += 1
+
+
+@pytest.mark.asyncio
+async def test_prompt_defaults_are_loaded_without_override():
+    db = _PromptSettingsDB()
+
+    prompt = await get_effective_prompt(db, "earth.news.enrich")
+
+    assert prompt.key == "earth.news.enrich"
+    assert prompt.is_custom is False
+    assert "strict JSON" in prompt.prompt
+
+
+@pytest.mark.asyncio
+async def test_prompt_override_save_and_reset():
+    db = _PromptSettingsDB()
+
+    saved = await save_prompt_override(
+        db,
+        "alerts.brief",
+        system_prompt="system custom",
+        prompt="prompt custom",
+    )
+
+    assert saved.is_custom is True
+    assert saved.system_prompt == "system custom"
+    assert saved.prompt == "prompt custom"
+    assert db.commits == 1
+
+    effective = await get_effective_prompt(db, "alerts.brief")
+    assert effective.prompt == "prompt custom"
+
+    reset = await reset_prompt_override(db, "alerts.brief")
+    assert reset.is_custom is False
+    assert reset.prompt != "prompt custom"
+
+
+@pytest.mark.asyncio
+async def test_prompt_list_marks_custom_items():
+    db = _PromptSettingsDB(
+        {
+            "overrides": {
+                "bgp.brief": {
+                    "system_prompt": "",
+                    "prompt": "custom bgp prompt",
+                    "updated_at": "2026-05-16T00:00:00Z",
+                }
+            }
+        }
+    )
+
+    prompts = await list_effective_prompts(db)
+    by_key = {prompt.key: prompt for prompt in prompts}
+
+    assert by_key["bgp.brief"].is_custom is True
+    assert by_key["bgp.brief"].prompt == "custom bgp prompt"
+    assert by_key["earth.news.enrich"].is_custom is False
--- a/backend/tests/test_settings_secret_placeholder.py
+++ b/backend/tests/test_settings_secret_placeholder.py
@@ -0,0 +1,12 @@
+from app.api.v1.settings import _is_secret_placeholder
+
+
+def test_secret_placeholder_treats_masked_values_as_placeholder():
+    assert _is_secret_placeholder("••••••••", "••••1234") is True
+    assert _is_secret_placeholder("********", "") is True
+    assert _is_secret_placeholder("  * * *  ", "") is True
+
+
+def test_secret_placeholder_preserves_real_keys_with_asterisks():
+    assert _is_secret_placeholder("sk-live-*real-key*", "") is False
+    assert _is_secret_placeholder("token_with*embedded*star", "") is False
--- a/backend/tests/test_situational_alert_ai_brief.py
+++ b/backend/tests/test_situational_alert_ai_brief.py
@@ -0,0 +1,71 @@
+import pytest
+
+from app.services.situational_alert_ai_brief import build_situational_alert_brief_request
+
+
+class _SingleUseScalarResult:
+    def __init__(self, value=0, rows=None):
+        self.value = value
+        self.rows = rows or []
+        self.scalar_calls = 0
+
+    def scalar(self):
+        self.scalar_calls += 1
+        if self.scalar_calls > 1:
+            raise AssertionError("scalar result was consumed more than once")
+        return self.value
+
+    def fetchall(self):
+        return self.rows
+
+    def scalar_one_or_none(self):
+        return None
+
+    def scalars(self):
+        rows = self.rows
+
+        class _Scalars:
+            def all(self):
+                return rows
+
+        return _Scalars()
+
+
+class _FakeBriefSession:
+    def __init__(self):
+        self._results = [
+            _SingleUseScalarResult(3),
+            _SingleUseScalarResult(2),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(4),
+            _SingleUseScalarResult(1),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(5),
+            _SingleUseScalarResult(2),
+            _SingleUseScalarResult(rows=[]),
+            _SingleUseScalarResult(),
+        ]
+
+    async def execute(self, _query):
+        return self._results.pop(0)
+
+
+@pytest.mark.asyncio
+async def test_situational_alert_brief_builder_reuses_counts_without_reconsuming_results(monkeypatch):
+    monkeypatch.setattr(
+        "app.services.situational_alert_ai_brief.get_latest_bgp_brief_record",
+        lambda: None,
+    )
+
+    request, facts, context = await build_situational_alert_brief_request(_FakeBriefSession())
+
+    assert request.title == "态势告警 AI 简报"
+    assert "总告警 3 条，active 2 条" in facts[0]
+    assert "累计 incidents 4 条，active incidents 1 条" in facts[1]
+    assert "累计 anomalies 5 条，active anomalies 2 条" in facts[2]
+    assert context["active_system_alerts"] == 2
+    assert context["active_bgp_incidents"] == 1
+    assert context["active_bgp_anomalies"] == 2
--- a/backend/tests/test_vessels.py
+++ b/backend/tests/test_vessels.py
@@ -574,13 +574,12 @@ async def test_vessel_snapshot_filters_type_and_bbox(monkeypatch):


@pytest.mark.asyncio
-async def test_legacy_vessels_geojson_endpoint_is_gone():
+async def test_legacy_vessels_geojson_route_is_not_registered():
    transport = ASGITransport(app=app)
    async with AsyncClient(transport=transport, base_url="http://test") as client:
        response = await client.get("/api/v1/visualization/geo/vessels")

-    assert response.status_code == 410
-    assert "/api/v1/vessels/snapshot" in response.json()["detail"]
+    assert response.status_code == 404


@pytest.mark.asyncio
@@ -669,15 +668,49 @@ async def test_vessel_snapshot_filters_bbox_and_caps_limit(monkeypatch):


@pytest.mark.asyncio
-async def test_legacy_vessels_geojson_rejects_even_with_bbox():
-    transport = ASGITransport(app=app)
-    async with AsyncClient(transport=transport, base_url="http://test") as client:
-        response = await client.get(
-            "/api/v1/visualization/geo/vessels",
-            params={"bbox": "10,59,11,60", "type": "cargo", "limit": 1000},
-        )
+async def test_vessel_snapshot_uses_legacy_fallback_when_raw_window_is_empty(monkeypatch):
+    now = datetime(2026, 4, 28, 1, 0, tzinfo=timezone.utc)
+    monkeypatch.setattr(
+        visualization,
+        "get_aggregated_vessels_snapshot",
+        AsyncMock(return_value=[]),
+    )
+    monkeypatch.setattr(
+        visualization,
+        "_load_legacy_vessel_snapshot_features",
+        AsyncMock(
+            return_value=[
+                {
+                    "type": "Feature",
+                    "id": 257123000,
+                    "geometry": {"type": "Point", "coordinates": [10.73, 59.91]},
+                    "properties": {
+                        "mmsi": 257123000,
+                        "name": "OSLO TRADER",
+                        "vessel_type": 70,
+                        "vessel_type_name": "Cargo",
+                        "received_at": now.isoformat(),
+                    },
+                }
+            ]
+        ),
+    )

-    assert response.status_code == 410
+    result = await visualization.build_vessel_snapshot_response(
+        object(),
+        bbox=(10.0, 59.0, 11.0, 60.0),
+        zoom=12,
+        type_filter=None,
+        limit=1000,
+        since_minutes=60,
+    )
+
+    assert result["count"] == 1
+    assert result["features"][0]["properties"]["name"] == "OSLO TRADER"
+    assert result["diagnostics"]["raw_feature_count"] == 0
+    assert result["diagnostics"]["legacy_feature_count"] == 1
+    assert result["diagnostics"]["legacy_backfilled_mmsi"] == 1
+    assert result["diagnostics"]["legacy_fallback_used"] is True


@pytest.mark.asyncio
--- a/config/earth-boundary-pov-policy.china-v1.json
+++ b/config/earth-boundary-pov-policy.china-v1.json
@@ -0,0 +1,92 @@
+{
+  "schema": "planet-earth-boundary-pov-policy/v1",
+  "profile": "china-pov-v1",
+  "description": "Product boundary policy for the China POV Earth boundary build. This file declares intent only; geometry must come from audited source packages and be applied offline before PMTiles/MVT generation.",
+  "defaultCountryHandling": "source-admin0-with-reviewed-overrides",
+  "rules": [
+    {
+      "id": "china-zangnan",
+      "name": "Zangnan / South Tibet",
+      "action": "union_to_country",
+      "targetIsoA3": "CHN",
+      "subtractFromIsoA3": ["IND"],
+      "hoverIsoA3": "CHN",
+      "labelPolicy": "show_country_only"
+    },
+    {
+      "id": "china-aksai-chin",
+      "name": "Aksai Chin",
+      "action": "union_to_country",
+      "targetIsoA3": "CHN",
+      "subtractFromIsoA3": ["IND"],
+      "hoverIsoA3": "CHN",
+      "labelPolicy": "show_country_only"
+    },
+    {
+      "id": "china-taiwan-penghu",
+      "name": "Taiwan and Penghu",
+      "action": "union_to_country",
+      "targetIsoA3": "CHN",
+      "hoverIsoA3": "CHN",
+      "labelPolicy": "show_country_only"
+    },
+    {
+      "id": "china-diaoyu-chiwei",
+      "name": "Diaoyu Dao, affiliated islands, and Chiwei Yu",
+      "action": "union_to_country",
+      "targetIsoA3": "CHN",
+      "hoverIsoA3": "CHN",
+      "labelPolicy": "show_country_only"
+    },
+    {
+      "id": "china-south-china-sea-islands",
+      "name": "Dongsha, Xisha, Zhongsha, Nansha, Huangyan Dao, Zengmu Ansha and related islands/reefs",
+      "action": "union_to_country",
+      "targetIsoA3": "CHN",
+      "hoverIsoA3": "CHN",
+      "labelPolicy": "show_country_only"
+    },
+    {
+      "id": "china-maritime-claim-line",
+      "name": "South China Sea dashed maritime claim line",
+      "action": "render_claim_line",
+      "targetIsoA3": "CHN",
+      "geometryRole": "claim_line_only",
+      "landPolygonEffect": "none"
+    },
+    {
+      "id": "kosovo",
+      "name": "Kosovo",
+      "action": "render_as_disputed_with_parent",
+      "parentIsoA3": "SRB",
+      "hoverIsoA3": "SRB",
+      "boundaryStyle": "disputed_internal",
+      "labelPolicy": "show_parent_country"
+    },
+    {
+      "id": "gaza",
+      "name": "Gaza Strip",
+      "action": "render_as_region_of_country",
+      "targetIsoA3": "PSE",
+      "hoverIsoA3": "PSE",
+      "boundaryStyle": "admin_or_disputed",
+      "labelPolicy": "show_country_only"
+    }
+  ],
+  "sourceRequirements": {
+    "geometryMustBeAudited": true,
+    "noHandDrawnClaimLines": true,
+    "noFrontendRuntimePovPatch": true,
+    "artifactIsolation": "one PMTiles/MVT artifact per POV profile"
+  },
+  "officialPositionNotes": [
+    {
+      "id": "kosovo",
+      "note": "China has emphasized respect for Serbia's sovereignty and territorial integrity and the framework of UNSC Resolution 1244."
+    },
+    {
+      "id": "gaza",
+      "note": "China supports the two-state solution and an independent State of Palestine based on the 1967 borders with East Jerusalem as its capital; Gaza governance should follow Palestinians governing Palestine."
+    }
+  ]
+}
--- a/config/earth-boundary-sources.example.json
+++ b/config/earth-boundary-sources.example.json
@@ -0,0 +1,112 @@
+{
+  "policy": {
+    "runtimeFetch": false,
+    "profile": "china-pov-v1",
+    "povPolicyPath": "config/earth-boundary-pov-policy.china-v1.json",
+    "productionTileFormat": "pmtiles+mvt",
+    "debugTileFormat": "geojson-directory",
+    "description": "Default Earth boundary update sources. These public Natural Earth endpoints make local high-precision boundary download work out of the box; replace with audited internal sources for production if needed."
+  },
+  "collectorConfigs": {
+    "earth_admin0_boundaries": {
+      "displayName": "Earth Admin-0 国界源",
+      "sourceKind": "admin0-boundaries",
+      "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_admin_0_countries.geojson",
+      "method": "GET",
+      "headers": {},
+      "auth_type": "none",
+      "license": "Natural Earth public domain",
+      "mapping_json": {
+        "source": {
+          "items_path": "$.features[*]"
+        },
+        "fields": {
+          "source_id": {
+            "path": "$.properties.id",
+            "type": "string"
+          },
+          "name": {
+            "path": "$.properties.name",
+            "type": "string"
+          },
+          "geometry": {
+            "path": "$.geometry",
+            "type": "object"
+          },
+          "properties": {
+            "path": "$.properties",
+            "type": "object"
+          }
+        }
+      }
+    },
+    "earth_coastline": {
+      "displayName": "Earth 海岸线源",
+      "sourceKind": "coastline",
+      "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_coastline.geojson",
+      "method": "GET",
+      "headers": {},
+      "auth_type": "none",
+      "license": "Natural Earth public domain",
+      "mapping_json": {
+        "source": {
+          "items_path": "$.features[*]"
+        },
+        "fields": {
+          "source_id": {
+            "path": "$.properties.id",
+            "type": "string"
+          },
+          "name": {
+            "path": "$.properties.name",
+            "type": "string"
+          },
+          "geometry": {
+            "path": "$.geometry",
+            "type": "object"
+          },
+          "properties": {
+            "path": "$.properties",
+            "type": "object"
+          }
+        }
+      }
+    },
+    "earth_claim_lines": {
+      "displayName": "Earth 主张线源",
+      "sourceKind": "claim-lines",
+      "endpoint": "https://raw.githubusercontent.com/nvkelso/natural-earth-vector/master/geojson/ne_10m_admin_0_boundary_lines_disputed_areas.geojson",
+      "method": "GET",
+      "headers": {},
+      "auth_type": "none",
+      "license": "Natural Earth public domain",
+      "mapping_json": {
+        "source": {
+          "items_path": "$.features[*]"
+        },
+        "fields": {
+          "source_id": {
+            "path": "$.properties.id",
+            "type": "string"
+          },
+          "name": {
+            "path": "$.properties.name",
+            "type": "string"
+          },
+          "geometry": {
+            "path": "$.geometry",
+            "type": "object"
+          },
+          "properties": {
+            "path": "$.properties",
+            "type": "object"
+          }
+        }
+      }
+    }
+  },
+  "notes": [
+    "Earth can download these sources directly from the toolbar settings when no local source override exists.",
+    "If tippecanoe/pmtiles are unavailable, the backend generates a GeoJSON high-precision package so the feature remains usable."
+  ]
+}
--- a/deploy/helm/planet/Chart.yaml
+++ b/deploy/helm/planet/Chart.yaml
@@ -0,0 +1,6 @@
+apiVersion: v2
+name: planet
+description: Planet situational awareness platform
+type: application
+version: 0.1.0
+appVersion: "0.52.0"
--- a/deploy/helm/planet/templates/_helpers.tpl
+++ b/deploy/helm/planet/templates/_helpers.tpl
@@ -0,0 +1,33 @@
+{{- define "planet.name" -}}
+{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+
+{{- define "planet.fullname" -}}
+{{- if .Values.fullnameOverride -}}
+{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" -}}
+{{- else -}}
+{{- $name := include "planet.name" . -}}
+{{- if contains $name .Release.Name -}}
+{{- .Release.Name | trunc 63 | trimSuffix "-" -}}
+{{- else -}}
+{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+{{- end -}}
+{{- end -}}
+
+{{- define "planet.labels" -}}
+app.kubernetes.io/name: {{ include "planet.name" . }}
+app.kubernetes.io/instance: {{ .Release.Name }}
+app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
+app.kubernetes.io/managed-by: {{ .Release.Service }}
+helm.sh/chart: {{ .Chart.Name }}-{{ .Chart.Version | replace "+" "_" }}
+{{- end -}}
+
+{{- define "planet.selectorLabels" -}}
+app.kubernetes.io/name: {{ include "planet.name" . }}
+app.kubernetes.io/instance: {{ .Release.Name }}
+{{- end -}}
+
+{{- define "planet.image" -}}
+{{- printf "%s/%s/%s:%s" .root.Values.global.imageRegistry .root.Values.global.imageNamespace .repository .root.Values.image.tag -}}
+{{- end -}}
--- a/deploy/helm/planet/templates/aiprovider.yaml
+++ b/deploy/helm/planet/templates/aiprovider.yaml
@@ -0,0 +1,66 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: {{ include "planet.fullname" . }}-aiprovider
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: aiprovider
+spec:
+  replicas: {{ .Values.aiprovider.replicaCount }}
+  selector:
+    matchLabels:
+      {{- include "planet.selectorLabels" . | nindent 6 }}
+      app.kubernetes.io/component: aiprovider
+  template:
+    metadata:
+      labels:
+        {{- include "planet.selectorLabels" . | nindent 8 }}
+        app.kubernetes.io/component: aiprovider
+    spec:
+      {{- with .Values.global.imagePullSecrets }}
+      imagePullSecrets:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      containers:
+        - name: aiprovider
+          image: {{ include "planet.image" (dict "root" . "repository" .Values.aiprovider.image.repository) | quote }}
+          imagePullPolicy: {{ .Values.image.pullPolicy }}
+          ports:
+            - name: http
+              containerPort: 8010
+          envFrom:
+            - configMapRef:
+                name: {{ include "planet.fullname" . }}-config
+            - secretRef:
+                name: {{ include "planet.fullname" . }}-secrets
+          readinessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 10
+            periodSeconds: 10
+          livenessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 30
+            periodSeconds: 20
+          resources:
+            {{- toYaml .Values.aiprovider.resources | nindent 12 }}
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "planet.fullname" . }}-aiprovider
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: aiprovider
+spec:
+  type: ClusterIP
+  ports:
+    - name: http
+      port: {{ .Values.aiprovider.service.port }}
+      targetPort: http
+  selector:
+    {{- include "planet.selectorLabels" . | nindent 4 }}
+    app.kubernetes.io/component: aiprovider
--- a/deploy/helm/planet/templates/backend.yaml
+++ b/deploy/helm/planet/templates/backend.yaml
@@ -0,0 +1,66 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: {{ include "planet.fullname" . }}-backend
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: backend
+spec:
+  replicas: {{ .Values.backend.replicaCount }}
+  selector:
+    matchLabels:
+      {{- include "planet.selectorLabels" . | nindent 6 }}
+      app.kubernetes.io/component: backend
+  template:
+    metadata:
+      labels:
+        {{- include "planet.selectorLabels" . | nindent 8 }}
+        app.kubernetes.io/component: backend
+    spec:
+      {{- with .Values.global.imagePullSecrets }}
+      imagePullSecrets:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      containers:
+        - name: backend
+          image: {{ include "planet.image" (dict "root" . "repository" .Values.backend.image.repository) | quote }}
+          imagePullPolicy: {{ .Values.image.pullPolicy }}
+          ports:
+            - name: http
+              containerPort: 8000
+          envFrom:
+            - configMapRef:
+                name: {{ include "planet.fullname" . }}-config
+            - secretRef:
+                name: {{ include "planet.fullname" . }}-secrets
+          readinessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 10
+            periodSeconds: 10
+          livenessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 30
+            periodSeconds: 20
+          resources:
+            {{- toYaml .Values.backend.resources | nindent 12 }}
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "planet.fullname" . }}-backend
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: backend
+spec:
+  type: ClusterIP
+  ports:
+    - name: http
+      port: {{ .Values.backend.service.port }}
+      targetPort: http
+  selector:
+    {{- include "planet.selectorLabels" . | nindent 4 }}
+    app.kubernetes.io/component: backend
--- a/deploy/helm/planet/templates/configmap.yaml
+++ b/deploy/helm/planet/templates/configmap.yaml
@@ -0,0 +1,27 @@
+{{- $postgresHost := .Values.postgresql.external.host -}}
+{{- if .Values.postgresql.internal.enabled -}}
+{{- $postgresHost = printf "%s-postgresql" (include "planet.fullname" .) -}}
+{{- end -}}
+{{- $redisHost := .Values.redis.external.host -}}
+{{- if .Values.redis.internal.enabled -}}
+{{- $redisHost = printf "%s-redis" (include "planet.fullname" .) -}}
+{{- end -}}
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: {{ include "planet.fullname" . }}-config
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+data:
+  DATABASE_URL: "postgresql+asyncpg://{{ .Values.postgresql.external.username }}:{{ .Values.postgresql.external.password }}@{{ $postgresHost }}:{{ .Values.postgresql.external.port }}/{{ .Values.postgresql.external.database }}"
+  REDIS_SERVER: {{ $redisHost | quote }}
+  REDIS_PORT: {{ .Values.redis.external.port | quote }}
+  REDIS_DB: {{ .Values.redis.external.db | quote }}
+  AI_BASE_URL: "http://{{ include "planet.fullname" . }}-aiprovider:{{ .Values.aiprovider.service.port }}"
+  AI_PROVIDER_API: "http://{{ include "planet.fullname" . }}-aiprovider:{{ .Values.aiprovider.service.port }}"
+  {{- range $key, $value := .Values.backend.env }}
+  {{ $key }}: {{ $value | quote }}
+  {{- end }}
+  {{- range $key, $value := .Values.aiprovider.env }}
+  {{ $key }}: {{ $value | quote }}
+  {{- end }}
--- a/deploy/helm/planet/templates/frontend.yaml
+++ b/deploy/helm/planet/templates/frontend.yaml
@@ -0,0 +1,61 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: {{ include "planet.fullname" . }}-frontend
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: frontend
+spec:
+  replicas: {{ .Values.frontend.replicaCount }}
+  selector:
+    matchLabels:
+      {{- include "planet.selectorLabels" . | nindent 6 }}
+      app.kubernetes.io/component: frontend
+  template:
+    metadata:
+      labels:
+        {{- include "planet.selectorLabels" . | nindent 8 }}
+        app.kubernetes.io/component: frontend
+    spec:
+      {{- with .Values.global.imagePullSecrets }}
+      imagePullSecrets:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      containers:
+        - name: frontend
+          image: {{ include "planet.image" (dict "root" . "repository" .Values.frontend.image.repository) | quote }}
+          imagePullPolicy: {{ .Values.image.pullPolicy }}
+          ports:
+            - name: http
+              containerPort: 3000
+          readinessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 5
+            periodSeconds: 10
+          livenessProbe:
+            httpGet:
+              path: /health
+              port: http
+            initialDelaySeconds: 20
+            periodSeconds: 20
+          resources:
+            {{- toYaml .Values.frontend.resources | nindent 12 }}
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "planet.fullname" . }}-frontend
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: frontend
+spec:
+  type: ClusterIP
+  ports:
+    - name: http
+      port: {{ .Values.frontend.service.port }}
+      targetPort: http
+  selector:
+    {{- include "planet.selectorLabels" . | nindent 4 }}
+    app.kubernetes.io/component: frontend
--- a/deploy/helm/planet/templates/ingress.yaml
+++ b/deploy/helm/planet/templates/ingress.yaml
@@ -0,0 +1,29 @@
+{{- if .Values.frontend.ingress.enabled }}
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: {{ include "planet.fullname" . }}
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+  {{- with .Values.frontend.ingress.annotations }}
+  annotations:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}
+spec:
+  ingressClassName: {{ .Values.frontend.ingress.className | quote }}
+  {{- with .Values.frontend.ingress.tls }}
+  tls:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}
+  rules:
+    - host: {{ .Values.frontend.ingress.host | quote }}
+      http:
+        paths:
+          - path: /
+            pathType: Prefix
+            backend:
+              service:
+                name: {{ include "planet.fullname" . }}-frontend
+                port:
+                  name: http
+{{- end }}
--- a/deploy/helm/planet/templates/secrets.yaml
+++ b/deploy/helm/planet/templates/secrets.yaml
@@ -0,0 +1,14 @@
+apiVersion: v1
+kind: Secret
+metadata:
+  name: {{ include "planet.fullname" . }}-secrets
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+type: Opaque
+stringData:
+  {{- range $key, $value := .Values.backend.secretEnv }}
+  {{ $key }}: {{ $value | quote }}
+  {{- end }}
+  {{- range $key, $value := .Values.aiprovider.secretEnv }}
+  {{ $key }}: {{ $value | quote }}
+  {{- end }}
--- a/deploy/helm/planet/templates/single-node-deps.yaml
+++ b/deploy/helm/planet/templates/single-node-deps.yaml
@@ -0,0 +1,116 @@
+{{- if .Values.postgresql.internal.enabled }}
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: {{ include "planet.fullname" . }}-postgresql
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: postgresql
+spec:
+  serviceName: {{ include "planet.fullname" . }}-postgresql
+  replicas: 1
+  selector:
+    matchLabels:
+      {{- include "planet.selectorLabels" . | nindent 6 }}
+      app.kubernetes.io/component: postgresql
+  template:
+    metadata:
+      labels:
+        {{- include "planet.selectorLabels" . | nindent 8 }}
+        app.kubernetes.io/component: postgresql
+    spec:
+      containers:
+        - name: postgresql
+          image: postgres:15
+          ports:
+            - name: postgres
+              containerPort: 5432
+          env:
+            - name: POSTGRES_USER
+              value: {{ .Values.postgresql.external.username | quote }}
+            - name: POSTGRES_PASSWORD
+              value: {{ .Values.postgresql.external.password | quote }}
+            - name: POSTGRES_DB
+              value: {{ .Values.postgresql.external.database | quote }}
+          readinessProbe:
+            exec:
+              command: ["pg_isready", "-U", {{ .Values.postgresql.external.username | quote }}]
+            initialDelaySeconds: 5
+            periodSeconds: 10
+          volumeMounts:
+            - name: data
+              mountPath: /var/lib/postgresql/data
+  volumeClaimTemplates:
+    - metadata:
+        name: data
+      spec:
+        accessModes: ["ReadWriteOnce"]
+        resources:
+          requests:
+            storage: 8Gi
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "planet.fullname" . }}-postgresql
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: postgresql
+spec:
+  ports:
+    - name: postgres
+      port: 5432
+      targetPort: postgres
+  selector:
+    {{- include "planet.selectorLabels" . | nindent 4 }}
+    app.kubernetes.io/component: postgresql
+{{- end }}
+{{- if .Values.redis.internal.enabled }}
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: {{ include "planet.fullname" . }}-redis
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: redis
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      {{- include "planet.selectorLabels" . | nindent 6 }}
+      app.kubernetes.io/component: redis
+  template:
+    metadata:
+      labels:
+        {{- include "planet.selectorLabels" . | nindent 8 }}
+        app.kubernetes.io/component: redis
+    spec:
+      containers:
+        - name: redis
+          image: redis:7-alpine
+          ports:
+            - name: redis
+              containerPort: 6379
+          readinessProbe:
+            exec:
+              command: ["redis-cli", "ping"]
+            initialDelaySeconds: 5
+            periodSeconds: 10
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "planet.fullname" . }}-redis
+  labels:
+    {{- include "planet.labels" . | nindent 4 }}
+    app.kubernetes.io/component: redis
+spec:
+  ports:
+    - name: redis
+      port: 6379
+      targetPort: redis
+  selector:
+    {{- include "planet.selectorLabels" . | nindent 4 }}
+    app.kubernetes.io/component: redis
+{{- end }}
--- a/deploy/helm/planet/values.single-node.yaml
+++ b/deploy/helm/planet/values.single-node.yaml
@@ -0,0 +1,29 @@
+global:
+  imageRegistry: gitea.rclaw.top
+  imageNamespace: linkong/planet
+
+image:
+  tag: latest
+
+frontend:
+  ingress:
+    enabled: true
+    host: planet.local
+
+postgresql:
+  internal:
+    enabled: true
+  external:
+    host: planet-postgresql
+    port: 5432
+    database: planet_db
+    username: postgres
+    password: postgres
+
+redis:
+  internal:
+    enabled: true
+  external:
+    host: planet-redis
+    port: 6379
+    db: 0
--- a/deploy/helm/planet/values.yaml
+++ b/deploy/helm/planet/values.yaml
@@ -0,0 +1,73 @@
+global:
+  imageRegistry: gitea.rclaw.top
+  imageNamespace: linkong/planet
+  imagePullSecrets: []
+
+fullnameOverride: planet
+
+image:
+  tag: latest
+  pullPolicy: IfNotPresent
+
+frontend:
+  replicaCount: 1
+  image:
+    repository: frontend
+  service:
+    port: 3000
+  ingress:
+    enabled: true
+    className: nginx
+    host: planet.example.com
+    annotations: {}
+    tls: []
+  resources: {}
+
+backend:
+  replicaCount: 1
+  image:
+    repository: backend
+  service:
+    port: 8000
+  env:
+    PROJECT_NAME: Planet
+    CORS_ORIGINS: '["*"]'
+  secretEnv:
+    SECRET_KEY: change-me
+  resources: {}
+
+aiprovider:
+  replicaCount: 1
+  image:
+    repository: aiprovider
+  service:
+    port: 8010
+  env:
+    AI_PROVIDER: openai
+    AI_MODEL: gpt-4o-mini
+  secretEnv:
+    AI_API_KEY: ""
+  resources: {}
+
+postgresql:
+  internal:
+    enabled: false
+  external:
+    host: postgres.example.com
+    port: 5432
+    database: planet_db
+    username: postgres
+    password: postgres
+
+redis:
+  internal:
+    enabled: false
+  external:
+    host: redis.example.com
+    port: 6379
+    db: 0
+
+smoke:
+  frontendPath: /
+  backendHealthPath: /health
+  aiProviderHealthPath: /health
--- a/docs/CHANGELOG.md
+++ b/docs/CHANGELOG.md
@@ -8,6 +8,138 @@ This project follows the repository versioning rule:
 - `improvement` -> `+0.0.1`（bugfix + 小功能混合）
 - `bugfix` -> `+0.0.1`

+## [0.60.0] — 2026-05-17
+
+Released: 2026-05-17
+
+### Highlights
+- 新增 Earth 内容与国界运行体验：品牌内容配置、国界精度滑块、高精国界首次应用按钮和高精重载构建入口共同收口。
+- 完成新闻中文展示链路收敛：滚动新闻改用一句话摘要，新闻本地化状态避免未完成内容进入滚动展示，英文原文继续保留在数据层。
+- 重构 AI Provider 与提示词边界：`aiprovider` 保持纯净模型适配，业务提示词集中到后端 AI task 默认注册表并支持运维台覆盖和重置。
+
+### Added / Fixed / Improved
+- 新增 AI 设置提示词 tab，按功能入口选择、编辑、保存和重置默认提示词，并修复态势告警 / BGP 简报展示生成内容而非 prompt。
+- 更新 Earth 新闻、HUD、巡航、桌面详情和移动详情的中文显示策略，并优化新闻摘要提示词为新闻式一句话概要。
+- 补充 Agent Runtime、Earth LLM 指令、语音识别和多角色态势感知的详细计划，并同步 README 架构说明、用户手册、FAQ 与开发者文档。
+- 补齐相关后端测试，覆盖 Earth 品牌配置、AI 简报接口和态势告警 AI 简报生成链路。
+
+---
+
+## [0.59.0] — 2026-05-16
+
+Released: 2026-05-16
+
+### Highlights
+- 将 Earth 国界从采集器体系迁移为 Earth 静态资产，恢复低精 GeoJSON fallback，并新增 Earth 工具栏高精国界下载/构建进度与热应用。
+- 重组后台“运维与配置”：新增 Earth 内容与采集管理二级入口，电视直播、国界精度、采集器、采集调度各归其位，未接入模块以占位页呈现。
+- 新增 AI task prompt 覆盖管理，按稳定 task key 管理新闻汉化、告警研判、BGP 简报等业务提示词，避免全局 prompt 污染。
+
+### Added / Fixed / Improved
+- Earth 新闻锚点链路增加队列化 enrichment 状态、Redis Streams 后台精修和 WebSocket patch 语义，前端汉化/锚点策略更稳定。
+- 国界 hover 与 interactable tooltip 解耦，鼠标位于国家 polygon 内时保持国界高亮，同时卫星/船只/BGP 等对象仍可显示自身信息。
+- 新增 `/api/v1/earth/boundaries/*` 状态、配置、构建和进度接口，并在启动初始化中清理旧 boundary datasource/task/snapshot 历史入口。
+- 补齐中英文用户手册、FAQ、quickstart、运维手册和开发者上下文文档，明确用户 UI、运维操作和开发者稳定边界。
+
+---
+
+## [0.58.0] — 2026-05-15
+
+Released: 2026-05-15
+
+### Highlights
+- 新增 Earth 高精度国界 PMTiles/MVT 前端链路，移除旧低精度 GeoJSON 国界兜底，国界缺失时显式报错。
+- 将 Earth 边界数据拆成 Admin-0、coastline、claim-lines 三个标准源采集器，并把 `earth_boundary_tiles` 收口为下游 PMTiles 构建器。
+- 修复 Earth 远距缩放下海陆基座与高清贴图 z-fighting 导致的雪花/黑块闪烁，并记录地表多层 shell 的深度间距规则。
+
+### Added / Fixed / Improved
+- 新增 `earth_boundary_source` 目标 schema、China POV policy 配置、PMTiles readiness/build 脚本和 collector artifact 登记流程。
+- Earth 新闻巡航改为优先使用可缓存的目标地点解析队列，并补充媒体新闻归档采集器与回归测试。
+- 调整数据源列表任务状态展示，让 Earth 采集器失败/未就绪状态可见，不再表现为“未执行”。
+- 更新中英文采集器、数据源设置、Earth 图层顺序、运维 runbook、FAQ、规则和计划文档。
+
+---
+
+## [0.57.0] — 2026-05-14
+
+Released: 2026-05-14
+
+### Highlights
+- 新增 WSL `--allow-lan` 临时 Windows relay，保持本机 `localhost:3000` / `localhost:8000` 不变，同时用 Windows 局域网 IP 暴露相同端口。
+- 启动脚本会检测旧 `netsh interface portproxy` 冲突并请求管理员 PowerShell 清理，避免 `svchost.exe / iphlpsvc` 持久占用 3000/8000。
+- 修复 Vite CJS Node API deprecated warning，将前端配置迁移到 ESM，并让脚本按实际后端端口注入代理目标。
+
+### Added / Fixed / Improved
+- 新增 [scripts/windows-lan-relay.ps1](/home/ray/dev/linkong/planet/scripts/windows-lan-relay.ps1)，在 Windows 侧启动随 WSL 服务健康状态自动退出的 TCP relay。
+- `planet.sh --allow-lan` 仅在 WSL + PowerShell 可用时启用 Windows relay，并自动检查 Windows 防火墙规则；非 WSL 环境保持原有路径。
+- 更新中英文 README、FAQ 和运维文档，说明旧 portproxy 清理、UAC 防火墙授权、同端口 localhost/LAN 访问和故障恢复方式。
+
+---
+
+## [0.56.0] — 2026-05-13
+
+Released: 2026-05-13
+
+### Highlights
+- 修复 Earth 卫星 SGP4 坐标口径，当前点/短尾迹使用地固坐标，锁定预测轨道使用固定地球姿态下的闭合惯性轨道。
+- 优化真实高度压缩显示上限，将高轨显示控制在地球半径外约四分之一，保持 GEO/MEO/LEO 分层同时避免轨迹过远。
+- 统一 BGP 光晕与图标色调，并更新超算中心建筑图标和 Earth 新闻/HUD 面板体验。
+
+### Added / Fixed / Improved
+- 修正卫星详情卡近地点/远地点高度计算，避免把轨道半径误显示为离地高度。
+- 调整 BGP event / collector halo 的浅色派生规则，使红色事件、橙色活跃观测站和蓝色 idle 观测站保持各自色相。
+- 补充中英文用户手册、FAQ、Earth frontend context、render order、layer style reference 和相关计划文档。
+
+---
+
+## [0.55.0] — 2026-05-13
+
+Released: 2026-05-13
+
+### Highlights
+- 新增 Earth 卫星真实高度显示开关，默认按 TLE/SGP4 真实轨道高度压缩分层，关闭后恢复旧版同层球面。
+- 将卫星轨迹显示保留在设置面板，并让轨迹、预测轨道和锁定视觉跟随真实高度开关即时刷新。
+- 修复 Earth 设置面板 switch 圆点垂直居中问题，并同步更新用户手册、FAQ 和 Earth 开发者文档。
+
+### Added / Fixed / Improved
+- 新增 `satelliteRealAltitudeEnabled` 本地偏好和真实高度压缩参数，保留 TLE 缺失/传播失败时的 fallback 固定高度。
+- 更新卫星渲染高度、轨迹、预测轨道和设置持久化链路，避免同一轨迹混入两套高度模型。
+- 补充中英文 manual、quickstart、FAQ、Earth frontend context、render order 和 layer style reference。
+
+---
+
+## [0.54.0] — 2026-05-13
+
+Released: 2026-05-13
+
+### Highlights
+- 新增 Earth 轨迹显示设置，将卫星轨迹从图层列表迁移到桌面与移动设置面板，便于独立控制轨迹线显示。
+- 优化卫星静止呼吸与低缩放圆点闪烁参数，让卫星态势视觉节奏更柔和、层次更清晰。
+- 更新算力中心图标资产，统一 Earth 交互图标在高分辨率显示下的视觉规格。
+
+### Added / Fixed / Improved
+- 新增轨迹显示偏好持久化，并在卫星图层不可用时同步禁用轨迹设置入口。
+- 调整卫星呼吸、透明度和 twinkle 参数，降低闪烁频率并增强可读性。
+- 将轨迹控制状态同步到桌面设置、移动设置和图层卡片刷新流程。
+
+---
+
+## [0.53.0] — 2026-05-13
+
+Released: 2026-05-13
+
+### Highlights
+- 新增正式交付基线：Gitea Actions CI、镜像发布、staging 自动部署，以及 Kubernetes Helm chart。
+- 前端生产镜像改为 `vite build` 静态产物 + nginx 托管，后端与 AI Provider 镜像移除开发 reload 并加入容器健康检查。
+- 明确 `planet.sh` 只作为开发启动入口，生产环境交给 Kubernetes Service、Ingress、readiness/liveness probe 和 rollout 管理。
+
+### Added / Fixed / Improved
+- 新增 frontend/backend/aiprovider 镜像 build smoke、Helm lint/template、staging rollout 与 HTTP smoke test 流水线。
+- 新增 Helm values、single-node 演示依赖、ConfigMap/Secret 引用、ClusterIP 服务和 frontend Ingress。
+- 更新运维文档与交付计划，保留 Vite 生产构建路线，不新增 Webpack 双构建链，Electron 暂不进入主线。
+- 继续收口启动脚本风险：状态目录、健康检查端口默认值、PID 校验、端口诊断和开发环境跨平台边界说明。
+
+---
+
 ## [0.52.0] — 2026-05-12

 Released: 2026-05-12
--- a/docs/plans/README.md
+++ b/docs/plans/README.md
@@ -19,7 +19,7 @@
 - [earth-mobile-drawer-ui-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-mobile-drawer-ui-plan.md)
 - [earth-compute-center-bgp-style-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-compute-center-bgp-style-plan.md)
 - [earth-renderer-architecture-separation-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-renderer-architecture-separation-plan.md)
- [earth-country-boundary-overlay-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-country-boundary-overlay-plan.md)
+- [Earth 高精度国界静态瓦片计划](/home/ray/dev/linkong/planet/docs/plans/earth-high-precision-boundary-tiles-plan.md)
 - [earth-predicted-orbit-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-predicted-orbit-plan.md)
 - [earth-webgl-instancing-satellites-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-webgl-instancing-satellites-plan.md)
 - [earth-real-terrain-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-real-terrain-plan.md)
--- a/docs/plans/agents-earth-command-runtime-plan.md
+++ b/docs/plans/agents-earth-command-runtime-plan.md
@@ -0,0 +1,810 @@
+# Agent Runtime, Earth LLM Command, And Speech Entry Plan
+
+## Summary
+
+Build an auditable backend Agent Runtime and upgrade the existing Earth search panel into a combined search, AI command, and voice wake entry. The first version is a runtime foundation, not the full multi-role simulation product yet.
+
+Typical user goals:
+
+- In Earth, type "高亮所有北斗卫星" and have the system open the satellite layer and highlight matching Beidou satellites.
+- Type "中国大陆的算力中心" and have the system open the compute-center layer, match mainland China compute centers, and highlight them.
+- When microphone permission is granted, use a configurable wake word, then speak an Earth command.
+- Save every Earth AI command as an agent run so operators can review the original input, speech transcription, tool steps, entity matches, final action plan, and frontend execution result.
+
+Core boundaries:
+
+- `aiprovider` remains the model gateway. It must not own business tools, database access, Earth actions, or agent policy.
+- The backend owns agent orchestration, tools, evidence storage, permission policy, and proposal application.
+- Earth v1 executes visualization actions only. It does not mutate business data.
+- Speech recognition uses a provider-neutral ASR API first, defaulting to OpenAI/Whisper-compatible transcription APIs, with local `whisper.cpp` style providers as later adapters.
+
+## Architecture
+
+### Backend Agent Runtime
+
+Recommended module shape:
+
+```text
+backend/app/models/
+  agent_run.py
+  agent_step.py
+  agent_evidence.py
+  agent_proposal.py
+
+backend/app/schemas/
+  agents.py
+  speech.py
+
+backend/app/services/agents/
+  runtime.py
+  orchestrator.py
+  tool_protocol.py
+  tool_registry.py
+  policy.py
+  proposals.py
+  earth_command.py
+  entity_query.py
+  speech.py
+
+backend/app/api/v1/
+  agents.py
+```
+
+Database conventions should follow the current project style:
+
+- Use SQLAlchemy models.
+- Import new models from `init_db()`.
+- Let `Base.metadata.create_all` create tables.
+- Add required indexes with `CREATE INDEX IF NOT EXISTS`.
+- Do not introduce Alembic for this feature.
+
+### aiprovider Boundary
+
+`aiprovider` should continue to provide model transport only:
+
+- Keep the existing `/v1/analyze` endpoint.
+- If schemas are extended, only pass through model/provider request fields and normalize responses.
+- Do not add WebSearch, database queries, Earth entity lookup, or business action execution inside `aiprovider`.
+- If a provider does not support native tool calling, the backend must use JSON tool-call fallback.
+
+### Agent Tool Protocol
+
+The first version should support two protocols:
+
+- Default path: backend JSON tool-call loop.
+- Optional path: provider-native tools when the configured provider supports them.
+- Fallback path: if provider-native tools are unavailable or unstable, automatically use JSON tool-call.
+
+Example JSON tool call:
+
+```json
+{
+  "type": "tool_call",
+  "tool": "earth.find_entities",
+  "arguments": {
+    "domain": "satellites",
+    "filters": {
+      "constellation": "beidou"
+    },
+    "limit": 500
+  }
+}
+```
+
+Example final response:
+
+```json
+{
+  "type": "final",
+  "summary": "已找到并高亮北斗卫星。",
+  "result": {
+    "actions": []
+  }
+}
+```
+
+Hard limits:
+
+- One run may execute at most 8 tool steps by default.
+- One tool call times out after 30 seconds by default.
+- Only registered whitelist tools may run.
+- Tool arguments must pass Pydantic validation.
+- Illegal tools, invalid arguments, and denied actions must be recorded as step errors.
+- LLM output may not directly write business state. Writes are either proposals or backend-executed policy-approved actions.
+
+## Data Model
+
+### AgentRun
+
+Suggested fields:
+
+- `id`: integer primary key.
+- `public_id`: public string id.
+- `run_type`: `earth_command`, `situational_awareness`, `config_proposal`, or `diagnostic`.
+- `status`: `queued`, `running`, `waiting_approval`, `completed`, `failed`, or `stopped`.
+- `title`: short display title.
+- `objective`: text objective.
+- `input`: JSONB original input, including text and audio metadata.
+- `context`: JSONB run context.
+- `result_markdown`: final human-readable output.
+- `result_json`: structured output, including Earth action plans.
+- `provider`: nullable provider id.
+- `model`: nullable model id.
+- `request_id`: nullable propagated request id.
+- `created_by`: user id.
+- `created_at`, `updated_at`, `completed_at`.
+- `error`: nullable text error.
+
+### AgentStep
+
+Suggested fields:
+
+- `id`.
+- `run_id`.
+- `step_index`.
+- `step_type`: `llm`, `tool`, `policy`, `action`, or `transcription`.
+- `status`: `pending`, `running`, `completed`, `failed`, or `skipped`.
+- `name`: step name, for example `earth.find_entities`.
+- `input`: JSONB.
+- `output`: JSONB.
+- `error`: nullable text.
+- `started_at`, `completed_at`.
+- `duration_ms`.
+
+### AgentEvidence
+
+Suggested fields:
+
+- `id`.
+- `run_id`.
+- `step_id`: nullable.
+- `evidence_type`: `internal_record`, `web_search`, `web_fetch`, `entity_match`, or `transcription`.
+- `source`: source id.
+- `title`: display title.
+- `url`: nullable source URL.
+- `content`: text evidence content.
+- `content_hash`: nullable hash.
+- `metadata`: JSONB.
+- `retrieved_at`.
+
+### AgentProposal
+
+Suggested fields:
+
+- `id`.
+- `run_id`.
+- `proposal_type`: `datasource_config`, `ai_prompt`, or `external_integration`.
+- `status`: `pending`, `approved`, `applied`, `rejected`, or `failed`.
+- `risk_level`: `low`, `medium`, or `high`.
+- `target`: JSONB target descriptor.
+- `before_payload`: JSONB.
+- `after_payload`: JSONB.
+- `rationale`: text.
+- `policy_result`: JSONB.
+- `applied_by`: nullable user id.
+- `applied_at`: nullable timestamp.
+- `error`: nullable text.
+
+## Public APIs
+
+### Agent Runs
+
+Add:
+
+```text
+POST /api/v1/agents/runs
+GET /api/v1/agents/runs
+GET /api/v1/agents/runs/{run_id}
+POST /api/v1/agents/runs/{run_id}/stop
+POST /api/v1/agents/proposals/{proposal_id}/apply
+```
+
+### Earth Command
+
+Add:
+
+```text
+POST /api/v1/agents/earth/command
+```
+
+Request shape:
+
+```json
+{
+  "input_text": "高亮所有北斗卫星",
+  "source": "text",
+  "transcription_id": null,
+  "client_context": {
+    "visible_layers": ["satellites"],
+    "locale": "zh-CN",
+    "viewport": {
+      "is_mobile": false
+    }
+  }
+}
+```
+
+Response shape:
+
+```json
+{
+  "run_id": "agent_xxx",
+  "summary": "已找到并高亮北斗卫星。",
+  "actions": [
+    {
+      "type": "show_layer",
+      "layer": "satellites",
+      "enabled": true
+    },
+    {
+      "type": "highlight_entities",
+      "domain": "satellites",
+      "entity_ids": ["satellite:norad:12345"],
+      "style": {
+        "color": "#7dd3fc"
+      }
+    },
+    {
+      "type": "open_result_panel",
+      "title": "北斗卫星",
+      "items": []
+    }
+  ],
+  "matched_entities": [],
+  "confidence": 0.86,
+  "missing_data": []
+}
+```
+
+### Speech / ASR
+
+Add:
+
+```text
+POST /api/v1/agents/speech/transcriptions
+```
+
+Request should use multipart form data:
+
+- `file`: audio blob.
+- `language`: default `zh`.
+- `provider`: optional provider override.
+- `source`: default `earth_command`.
+
+Response shape:
+
+```json
+{
+  "text": "高亮所有北斗卫星",
+  "provider": "openai_compatible",
+  "model": "whisper-1",
+  "duration_ms": 1234,
+  "confidence": null,
+  "metadata": {}
+}
+```
+
+## Earth Action Plan
+
+### Allowed Action Types
+
+The first version may only return:
+
+```text
+show_layer
+highlight_entities
+filter_entities
+focus_view
+open_result_panel
+clear_highlight
+```
+
+### Allowed Domains
+
+The first version supports:
+
+```text
+satellites
+compute_centers
+bgp
+news
+vessels
+cables
+```
+
+### EarthAction Shape
+
+Example:
+
+```json
+{
+  "type": "highlight_entities",
+  "domain": "satellites",
+  "entity_ids": ["satellite:norad:12345"],
+  "style": {
+    "color": "#7dd3fc",
+    "mode": "glow"
+  },
+  "reason": "用户要求高亮北斗卫星"
+}
+```
+
+### Safety Rules
+
+The backend must validate action plans before returning them:
+
+- `type` must be in the allowed action list.
+- `domain` must be in the allowed domain list.
+- `entity_ids` must come from backend or current Earth candidate data. The model may not invent ids.
+- One highlight action should include at most 500 entities by default. If more match, return a truncation note in `missing_data` or `summary`.
+- `focus_view` must include valid coordinates, a valid region, or a matched entity.
+- No action may return executable JavaScript, arbitrary CSS, or arbitrary URLs to fetch.
+
+## Earth Entity Query
+
+Add a backend entity query service used by both deterministic resolvers and LLM tools.
+
+### Satellites
+
+Sources:
+
+- `/api/v1/visualization/geo/satellites`.
+- Current TLE collected data.
+- GeoJSON feature properties.
+
+Filters:
+
+- `constellation`: `beidou`, `gps`, `galileo`, `glonass`, `starlink`, `iridium`, `geo`, `leo`.
+- Name contains.
+- NORAD id.
+- Country/operator when present in data.
+- Orbital class when inferable from existing fields.
+
+Beidou matching:
+
+- Prefer `constellation == beidou`.
+- Then match names containing `BEIDOU`, `BDS`, `BEIDOU-`, or `北斗`.
+- Stable entity id format should be `satellite:norad:{norad_id}` when possible, with fallback `satellite:index:{index}`.
+
+### Compute Centers
+
+Sources:
+
+- `/api/v1/visualization/geo/compute-centers`.
+- Unified TOP500 and Epoch AI GPU GeoJSON properties.
+
+Filters:
+
+- Country/region.
+- `site_type`: `supercomputer` or `gpu_cluster`.
+- Source: `top500` or `epoch_ai_gpu`.
+- Name contains.
+- `needs_confirmation`.
+- `location_precision`.
+
+Mainland China matching:
+
+- Match country values such as `China`, `中国`, or `People's Republic of China`.
+- Exclude obvious non-mainland records when fields identify Hong Kong, Macau, or Taiwan.
+- If records do not expose enough region detail to separate mainland China from Hong Kong, Macau, or Taiwan, return a `missing_data` note and conservatively match `country=China`.
+
+### BGP, News, Vessels, And Cables
+
+First-version basic support:
+
+- BGP: severity, status, region, collector, prefix, ASN.
+- News: region, source, localized title, localized summary.
+- Vessels: vessel type, country/area, name, status.
+- Cables: cable name, landing point, country/region.
+
+## Earth Frontend Integration
+
+### Search Panel Merge
+
+Reuse the existing Earth search panel:
+
+- Default behavior remains normal local search.
+- Add an AI command state for natural-language commands.
+- Add a "use AI" command button.
+- Add a microphone button.
+- Show running status, result summary, and clear-highlight action.
+- Visually distinguish ordinary search results from AI action results.
+
+Natural-language routing:
+
+- If the user clicks the AI command button, always call the AI command endpoint.
+- If input contains action words such as `高亮`, `显示`, `找出`, `聚焦`, `打开`, `筛选`, `隐藏`, or `清除`, suggest AI command mode.
+- Short ordinary keywords continue to use local search.
+
+### Earth Action Executor
+
+Add:
+
+```text
+frontend/public/earth/js/earth-command-actions.js
+```
+
+Responsibilities:
+
+- Receive backend `EarthActionPlan`.
+- Toggle required layers.
+- Apply highlights and filters.
+- Focus view when requested.
+- Open or update the result panel.
+- Clear the previous AI command highlight.
+
+Execution order:
+
+1. Clear previous AI command highlights unless the plan is explicitly additive.
+2. Show required layers.
+3. Apply filters and highlights.
+4. Focus the view if requested.
+5. Open the result panel.
+6. Record the execution summary in UI state.
+
+### Satellite Highlight
+
+Existing `highlightRelatedSatellites(indices, color)` can be reused, but Earth command support needs an entity-id-to-index map.
+
+Add:
+
+- Satellite id map built from current satellite data.
+- Highlight by `entity_ids`.
+- Clear AI command satellite highlights without clearing manual locked selection.
+- Auto-enable the satellite layer before highlighting.
+
+### Compute Center Highlight
+
+Existing compute center markers support marker state, but need a dedicated batch AI highlight state.
+
+Add:
+
+- Batch set compute center markers to AI highlighted state.
+- Keep AI highlight compatible with hover and locked states.
+- Clear AI command compute highlights without clearing manual locked selection.
+- Auto-enable the compute-center layer before highlighting.
+
+### Result Panel
+
+First version can show results inside the merged search panel:
+
+- Title, for example `北斗卫星`.
+- Count, for example `已高亮 32 个对象`.
+- List with the first 20 entities.
+- Actions: clear highlight, rerun, view Agent Run.
+
+## Voice And Wake Word
+
+### ASR Configuration
+
+Add Speech/ASR under the AI settings tool tab.
+
+Fields:
+
+- `enabled`.
+- `provider`.
+- `base_url`.
+- `api_key`.
+- `model`.
+- `language`.
+- `timeout_seconds`.
+- `max_audio_size_mb`.
+
+First-version providers:
+
+- `openai_whisper`.
+- `openai_compatible`.
+- `local_whisper`.
+
+Recommended defaults:
+
+- `provider`: `openai_compatible`.
+- `model`: `whisper-1` or user-configured equivalent.
+- `language`: `zh`.
+
+Secret handling:
+
+- Use the existing masked secret pattern.
+- Do not allow agent proposals to write API keys.
+- Support environment fallback such as `ASR_API_KEY` and `OPENAI_API_KEY`.
+
+### Wake Word
+
+The wake word is configurable.
+
+Implementation rules:
+
+- Treat wake word as a browser-local preference in v1.
+- Store it in `localStorage`.
+- Default suggestion: `小星球`.
+- Provide a wake-word setting inside the Earth search/command panel.
+- Do not listen until the user explicitly enables voice wake.
+- Do not upload audio before wake.
+- After wake, record one instruction audio segment and upload it to backend ASR.
+
+Fallbacks:
+
+- Browser does not support continuous local recognition: fall back to click-to-record.
+- Microphone permission denied: show `麦克风权限未开启，仍可输入文字指令`.
+- ASR not configured: show `语音识别未配置`; text commands remain available.
+- User can disable wake listening and keep manual microphone recording.
+
+### Wake Word Technical Choice
+
+Do not require local Whisper in v1.
+
+Recommended path:
+
+- Use browser Web Speech API for local wake-word detection when available.
+- Fall back to click-to-record when unavailable.
+- Upload only the post-wake instruction audio to backend ASR.
+- Add local `whisper.cpp` streaming wake-word or ASR provider later.
+
+## Prompt Registry
+
+Add default prompt keys:
+
+```text
+agents.runtime.system
+agents.earth.command
+agents.situational.assessment
+agents.config.proposal
+agents.roles.network
+agents.roles.bgp
+agents.roles.platform_ops
+agents.roles.business_impact
+```
+
+`agents.earth.command` must require:
+
+- Strict JSON output only.
+- Use only provided candidate entities.
+- Do not invent objects.
+- Do not return arbitrary code.
+- Do not modify business data.
+- Return `clarification_needed` when intent is unclear.
+- Return action plans matching the schema.
+- Use Chinese summary for Chinese user input.
+
+## Policy
+
+### Tool Policy
+
+- The LLM can request tools, but the backend decides whether to execute.
+- Every tool must declare:
+  - name.
+  - description.
+  - input schema.
+  - output schema.
+  - permission.
+  - side-effect level.
+- First-version side-effect levels:
+  - `read`: can run directly.
+  - `proposal`: can only generate proposals.
+  - `write`: can only execute during proposal apply.
+
+### Proposal Policy
+
+First version supports configuration proposal application with role-based automatic/manual gating:
+
+- `super_admin` may enable automatic application for low-risk proposals.
+- Normal admins must manually confirm.
+- High-risk proposals always require manual confirmation.
+- Proposals containing secret fields are rejected.
+- Proposals failing schema validation are rejected.
+- Before and after payloads must be saved.
+- Failed applications must save errors.
+
+Low-risk scope:
+
+- Datasource endpoint/config non-secret fields.
+- AI prompt overrides.
+- External integration non-secret fields.
+
+Out of scope for v1:
+
+- Alert acknowledge/resolve.
+- Data deletion.
+- User permission changes.
+- Authentication configuration changes.
+- Database schema changes by agent.
+- Mutation of original Earth collected data.
+
+## Agent Operations UI
+
+Add an `Agent` page under `运维与配置`.
+
+Run list:
+
+- Status.
+- Type.
+- Title.
+- Creator.
+- Model.
+- Time.
+- Duration.
+
+Run detail:
+
+- Input.
+- Status timeline.
+- LLM steps.
+- Tool steps.
+- Evidence.
+- Final result.
+- Proposals.
+
+Proposal apply:
+
+- Before/after diff.
+- Risk level.
+- Policy result.
+- Apply button.
+- Reject button.
+
+Earth command run detail:
+
+- Original text or speech transcription.
+- Matched entities.
+- Action plan.
+- Link back to Earth or copy run id.
+
+## Frontend Types
+
+Add or extend:
+
+```ts
+interface AgentRun {}
+interface AgentStep {}
+interface AgentEvidence {}
+interface AgentProposal {}
+interface EarthCommandRequest {}
+interface EarthActionPlan {}
+interface EarthAction {}
+interface SpeechTranscriptionResponse {}
+```
+
+Earth action executor API:
+
+```js
+executeEarthActionPlan(plan, context)
+clearEarthCommandHighlights()
+getEarthCommandExecutionState()
+```
+
+## Testing
+
+### Backend Tests
+
+Use `uv`.
+
+Suggested files:
+
+```text
+backend/tests/test_agents_runtime.py
+backend/tests/test_agent_tool_protocol.py
+backend/tests/test_agent_policy.py
+backend/tests/test_earth_agent_command.py
+backend/tests/test_speech_transcription.py
+```
+
+Coverage:
+
+- Create run.
+- Persist ordered run steps.
+- Parse valid JSON tool calls.
+- Reject invalid JSON tool calls without executing tools.
+- Reject unregistered tools.
+- Reject invalid tool arguments.
+- Pass through provider-native tools fields when configured.
+- Fall back to JSON tool calls when native tools are unavailable.
+- When WebSearch is not configured, record `missing_data` and do not fail Earth command.
+- "高亮所有北斗卫星" returns a satellite highlight action.
+- "中国大陆的算力中心" returns a compute-center highlight action.
+- Mainland China matching excludes Hong Kong, Macau, and Taiwan when fields permit it.
+- Field granularity limitations are returned in `missing_data`.
+- ASR not configured returns a clear error.
+- ASR provider success returns transcription.
+- Normal admins cannot auto-apply proposals.
+- `super_admin` can apply low-risk proposals.
+- Secret-field proposals are rejected.
+
+### Frontend Tests
+
+Use `bun`.
+
+Required build check:
+
+```bash
+cd frontend && bun run build
+```
+
+Suggested Earth JS tests:
+
+- Action executor opens the satellite layer.
+- Action executor highlights satellite entity ids.
+- Action executor highlights compute-center entity ids.
+- Clear highlight does not clear manual locked selection.
+- Search panel ordinary search still uses local search.
+- AI command button calls `/api/v1/agents/earth/command`.
+- ASR missing configuration and microphone permission denial show clear UI feedback.
+
+### Manual Acceptance
+
+1. Type `高亮所有北斗卫星`.
+   - Satellite layer opens.
+   - Beidou satellites are highlighted.
+   - Panel shows count and summary.
+   - Agent run is reviewable.
+
+2. Type `显示中国大陆的算力中心`.
+   - Compute-center layer opens.
+   - Mainland China related compute centers are highlighted.
+   - If data cannot separate mainland China from Hong Kong/Macau/Taiwan, the UI shows a missing-data note.
+
+3. Click microphone.
+   - Permission denial is clear.
+   - Permission grant allows recording.
+   - Configured ASR transcribes and executes the command.
+   - If ASR is not configured, text input still works.
+
+4. Enable voice wake.
+   - Wake word is configurable.
+   - Audio is not uploaded before wake.
+   - After wake, one instruction audio segment is uploaded.
+   - Wake listening can be disabled.
+
+5. Open Agent operations UI.
+   - Earth command run is listed.
+   - Transcription step is visible when speech was used.
+   - Entity query step is visible.
+   - Final action plan is visible.
+   - User can return to Earth or copy the run id.
+
+## Documentation
+
+When implemented, update:
+
+```text
+docs/technical/zh/agents-aiprovider.md
+docs/technical/en/agents-aiprovider.md
+docs/technical/zh/manual.md
+docs/technical/en/manual.md
+docs/technical/zh/earth-frontend-context.md
+docs/technical/en/earth-frontend-context.md
+```
+
+Document:
+
+- `aiprovider` and backend agent boundaries.
+- Earth AI command entry.
+- Speech/ASR configuration.
+- Wake-word privacy behavior.
+- Agent run/evidence/proposal review.
+- V1 capability boundary: Earth visualization actions only, no business data mutation.
+
+## Implementation Order
+
+1. Add backend Agent data models, `init_db()` imports, and indexes.
+2. Add Agent schemas and run CRUD API.
+3. Add tool registry, JSON tool-call parser, and policy skeleton.
+4. Add Earth entity query service.
+5. Add `agents.earth.command` prompt and backend command endpoint.
+6. Add frontend Earth action executor.
+7. Merge AI command entry into Earth search panel.
+8. Add satellite and compute-center batch highlight support.
+9. Add Speech/ASR settings and transcription API.
+10. Add Earth microphone recording, wake-word local setting, and fallback behavior.
+11. Add Agent operations UI.
+12. Add proposal apply policy and low-risk configuration application.
+13. Add tests and documentation updates.
+
+## Assumptions
+
+- Earth LLM v1 only executes visualization actions.
+- Earth commands create agent runs but do not automatically create configuration proposals.
+- Configuration proposals are applied only from the Agent operations UI.
+- Wake word is a device-local preference in v1 and is stored in `localStorage`.
+- ASR is API-first and Whisper-compatible by default; local Whisper is a later provider.
+- Multi-role simulation only reserves schemas and prompt keys in v1.
+- Palantir-style situational workflows should grow from the shared evidence, entity, action, and assessment model instead of a separate isolated system.
--- a/docs/plans/ai-prompt-settings-task-registry-plan.md
+++ b/docs/plans/ai-prompt-settings-task-registry-plan.md
@@ -0,0 +1,52 @@
+# AI Prompt Settings and Task Registry Plan
+
+## Summary
+
+Add an AI prompt settings tab under the operations AI settings page. Operators can select a business AI task from a dropdown, edit its prompt, save the override, and reset it back to the shipped default. Runtime LLM calls must resolve prompts through a task registry instead of embedding large prompt blocks at each call site.
+
+Default prompts are shipped as versioned resource data, not scattered business-code literals. Business services reference stable task keys, and the runtime resolves the effective prompt from the database override first, then the shipped default resource.
+
+## Key Changes
+
+- Add a backend task prompt registry with stable keys, labels, groups, versions, default system prompts, and default task prompts.
+- Store operator overrides in the existing `SystemSetting` table under an `ai_prompts` category. Store only custom overrides; defaults remain in the versioned prompt resource.
+- Add settings APIs:
+  - `GET /api/v1/settings/ai-prompts`
+  - `PUT /api/v1/settings/ai-prompts/{task_key}`
+  - `POST /api/v1/settings/ai-prompts/{task_key}/reset`
+- Migrate business LLM entrypoints to resolve prompts by task key. `aiprovider` remains a pure model adapter and does not inject business prompts.
+- Add a “提示词” tab to `/ai`. The tab shows a grouped task dropdown, current/default prompt status, editable prompt fields, save, and reset-to-default controls.
+
+## Initial Tasks
+
+- `earth.news.enrich` — Earth news localization and location enrichment.
+- `alerts.brief` — system alert AI brief.
+- `alerts.situational.brief` — situational alert AI brief.
+- `bgp.brief` — BGP AI brief.
+- `location.factcheck.normalize` — location factcheck normalization.
+- `location.factcheck.resolve` — location factcheck fallback resolution.
+- `datasource.mapping` — datasource mapping DSL generation.
+- `credential.guide` — credential guide generation.
+- `ai.connection_test` — AI provider connection test.
+
+Playground and public free-form analyze endpoints stay caller-controlled and are not shown in the prompt settings dropdown.
+
+## Test Plan
+
+- Backend uses `uv`:
+  - `uv run pytest tests/test_settings_ai_prompts.py`
+  - `uv run pytest tests/test_earth_news.py`
+  - `uv run pytest tests/test_api.py`
+- Frontend uses `bun`:
+  - `cd frontend && bun run build`
+- Manual checks:
+  - Prompt dropdown switches task content correctly.
+  - Save persists an override and runtime calls use it.
+  - Reset deletes the override and restores the shipped default.
+  - Alert prompts do not leak into news, BGP, datasource, or location tasks.
+
+## Assumptions
+
+- This iteration does not add prompt history, approval workflows, A/B testing, or per-user prompt variants.
+- Strict JSON tasks may fail validation if an operator edits away the output contract; existing task-specific failure and retry behavior remains responsible for recovery.
+- The UI stays Chinese-only in this iteration.
--- a/docs/plans/datasource-custom-api-mapping-plan.md
+++ b/docs/plans/datasource-custom-api-mapping-plan.md
@@ -13,7 +13,7 @@
 | 采集时是否调用 LLM | 不调用；采集链路必须确定性、可审计、可复现 |
 | 自定义数据如何进入 Earth | 必须映射到已支持的目标 schema，或先进入通用数据沉淀 |
 | 外部凭证放置位置 | Settings / 外部集成统一管理 provider token；DataSources 引用 provider profile |
-| TimescaleDB | 放入 TODO；高频时序数据稳定后再评估迁移 |
+| TimescaleDB | 高频时序数据稳定后再评估迁移 |

 ---

@@ -57,7 +57,7 @@ flowchart LR
 | schema | 用途 | Earth 可视化 |
 |-------|------|-------------|
 | `vessel_ais` | 船只 AIS 位置、航速、航向、MMSI 等 | 进入船舶图层 |
-| `geo_points` | 通用点位数据，包含经纬度、名称、类型、时间 | 进入通用 geo layer（TODO） |
+| `geo_points` | 通用点位数据，包含经纬度、名称、类型、时间 | 未来进入通用 geo layer |
 | `news_events` | 新闻/事件类数据，带时间、地点、摘要、来源 | 复用新闻/事件链路 |
 | `compute_centers` | 算力中心、机房、数据中心数据 | 复用算力中心图层 |
 | `generic_records` | 未知结构化数据沉淀 | 不直接展示 |
@@ -294,7 +294,7 @@ PostgreSQL 可以承载当前规模的采集、关系查询、JSONB 沉淀和基
 - 查询模式还没稳定。
 - 需要快速迭代 schema 与 mapping。

-### TODO：TimescaleDB
+### TimescaleDB 后续评估

 以下条件满足后，再评估 TimescaleDB：

@@ -324,7 +324,7 @@ PostgreSQL 可以承载当前规模的采集、关系查询、JSONB 沉淀和基
 - Settings 中保存 provider credentials。
 - API 返回配置时必须 mask secret。
 - LLM prompt 只能包含脱敏 sample 和 schema 说明。
- 后续 TODO：引入字段级加密或 KMS。
+- 后续可引入字段级加密或 KMS。

 ### Mapping 治理

--- a/docs/plans/docs-audience-split-plan.md
+++ b/docs/plans/docs-audience-split-plan.md
@@ -31,7 +31,7 @@
 3. **登录与找回密码** — 登录页、忘记密码流程
 4. **账户设置** — 修改密码、修改邮箱（需重新验证）、查看权限组、登出
 5. **Console 总览** — 左侧菜单结构、各路由用途
-6. **配置数据采集器** — `/settings?tab=collector_credentials`：选择 collector、连接测试、保存凭证；BarentsWatch / AISStream 两个典型例子
+6. **配置数据采集器** — `/collection-management?tab=collector_credentials`：选择 collector、连接测试、保存凭证；BarentsWatch / AISStream 两个典型例子
 7. **配置 AI 凭证** — `/ai?tab=providers`：默认 provider、模型、Base URL、API Key、本地代理；工具 tab（WebSearch、OCR）
 8. **系统设置** — `/settings` 其他子 tab（系统设置、电视直播源、SMTP 邮件）
 9. **用户管理（管理员）** — `/users`：创建、删除、改角色、Gatekeeper 权限组
@@ -48,7 +48,7 @@

 - 打开管理员给你的 URL
 - 注册账号 + 邮箱验证
- 登录后第一次做什么（建议先到 `/settings?tab=collector_credentials` 配一个 collector，再到 `/ai` 配模型）
+- 登录后第一次做什么（建议先到 `/collection-management?tab=collector_credentials` 配一个 collector，再到 `/ai` 配模型）
 - 看 Earth

 部署/开发的 quickstart 内容并入 `ops-runbook.md` 的"首次部署"小节，**不**再单独出 `ops-quickstart.md`，避免新增维护点。
--- a/docs/plans/earth-high-precision-boundary-tiles-plan.md
+++ b/docs/plans/earth-high-precision-boundary-tiles-plan.md
@@ -0,0 +1,71 @@
+# Earth High Precision Boundary PMTiles Plan
+
+## Status
+
+Superseded status:
+
+This plan originally treated boundaries as collector-managed source records. The current implementation has moved country boundaries out of the datasource / collector lifecycle. Boundaries are now Earth static rendering assets managed by `Operations and Configuration -> Earth Content -> Boundary Precision` and `/api/v1/earth/boundaries/*`. The bundled low-precision GeoJSON is the default fallback, and high precision is an opt-in local PMTiles build.
+
+Historical implementation notes below are retained only as context and must not be used as the current architecture:
+
+- Three standard source collectors now handle real source ingestion: `earth_admin0_boundaries`, `earth_coastline`, and `earth_claim_lines`.
+- Each source collector reads endpoint / headers / auth / `target_schema=earth_boundary_source` from Collector Settings, downloads the configured payload, stores the full artifact under `data/earth-boundary-sources/<collector>/<sha256>.*`, and writes a hash / feature-count / artifact-path record to `CollectedData`.
+- The backend `earth_boundary_tiles` item is now a downstream PMTiles builder. It refuses to run until the three source records exist, then skips rebuilds when source / POV policy / build config are unchanged.
+- The frontend boundary layer now requires the production `pmtiles-mvt` provider and no longer falls back to legacy low-precision GeoJSON.
+- Generated loose boundary data is ignored by Git and is not the production deployment format.
+- Production PMTiles builds require external `tippecanoe` and `pmtiles` CLIs; missing tools fail the builder clearly instead of registering fake tile records.
+
+Still required before claiming true one-to-one high precision:
+
+- Replace the repository seed GeoJSON with audited high-precision admin boundary, coastline, and claim-line source packages.
+- Run a real geometry preparation step that applies the China POV policy through union / subtract / validity repair before PMTiles creation.
+- Build and publish `earth-boundaries-china-pov-v1.pmtiles` plus its manifest.
+
+## Summary
+
+The Earth boundary layer should use one static PMTiles archive containing MVT tiles instead of thousands of loose GeoJSON files. The artifact is POV-specific: `earth-boundaries-china-pov-v1.pmtiles` has China POV baked in during offline source preparation, and the browser never patches political boundaries at runtime.
+
+Production should serve a single PMTiles artifact through static hosting and HTTP range requests. In development or on machines that have not opted into high precision, missing PMTiles falls back to the bundled low-precision GeoJSON so the Earth base remains usable.
+
+## Key Implementation Rules
+
+- Source inputs must be auditable. OSM admin boundaries, coastline packages, and claim-line endpoints are configured through Earth Content boundary precision settings; `config/earth-boundary-sources.example.json` remains the versioned example template.
+- China POV geometry is applied before tiling:
+  - Zangnan and Aksai Chin are unioned into China and subtracted from India.
+  - Taiwan/Penghu, Diaoyu Dao and affiliated islands, Chiwei Yu, and South China Sea islands are China hover/country features.
+  - The South China Sea dashed line is a claim-line layer only; it never consumes Malaysian, Philippine, Vietnamese, or other land polygons.
+  - Kosovo is not an independent country surface in this profile; Gaza is a Palestine region.
+- PMTiles/MVT layer names are fixed for the frontend:
+  - `boundary_admin0`
+  - `boundary_disputed_internal`
+  - `coastline`
+  - `claim_line`
+- The frontend provider is selected from local high-precision preference plus the boundary manifest:
+  - `tileProvider: "pmtiles-mvt"` reads the PMTiles artifact.
+  - Missing high-precision preference, missing manifest, or missing PMTiles artifact falls back to low-precision GeoJSON.
+- Redis is not part of v1. Static PMTiles plus browser/CDN range caching is the default performance model.
+
+## Cleanup And Documentation
+
+- Do not commit generated loose tiles under `frontend/public/earth/data/boundaries/` or source downloads under `data/earth-boundary-sources/`.
+- Remove stale generated debug data before production builds; regenerate it only when smoke testing the debug path.
+- Keep the high-level plan, backend collector docs, layer style docs, and ops runbook aligned whenever the provider contract changes.
+- After implementation changes, provide user-facing operation steps covering source configuration, artifact build/deploy, page verification, and fallback troubleshooting.
+
+## Verification
+
+- The Earth boundary build API reports missing source configuration or missing tools clearly, without creating datasource collection records.
+- The PMTiles builder fails as not ready when source artifacts exist but `tippecanoe` / `pmtiles` are missing.
+- Running the PMTiles builder twice returns `unchanged` on the second run when inputs are stable.
+- `git add . --dry-run` does not include generated loose boundary tiles or source downloads.
+- `/home/ray/.bun/bin/bun run build` passes in `frontend`.
+- Manual Earth checks confirm:
+  - PMTiles range requests are issued only for visible tiles.
+  - Boundary toggle, hover tooltip, and country highlight still work.
+  - PMTiles failure reports a high-precision boundary error; machines without high-precision enabled continue drawing low-precision fallback boundaries.
+
+## Assumptions
+
+- "One-to-one" means source-faithful to the selected audited vector source, not hand-tuned to a screenshot.
+- The China POV artifact is static and versioned; no runtime region-based POV switching is planned.
+- The repository low-precision seed file is retained as the runtime fallback for country boundaries.
--- a/docs/plans/earth-high-resolution-basemap-tiles-plan.md
+++ b/docs/plans/earth-high-resolution-basemap-tiles-plan.md
@@ -0,0 +1,118 @@
+# Earth High Resolution Basemap Tiles Plan
+
+## Summary
+
+High-precision borders now expose a separate visual problem: the vector coastline and border data are more accurate than the current raster Earth texture. The next step is a high-resolution basemap tile layer that aligns visually with the high-precision coastline instead of replacing the globe with one huge static image.
+
+Do not solve this by committing a larger single world texture. A single 16K/32K raster still wastes memory, loads slowly, and becomes blurry or misaligned when zooming into coastal detail. The target architecture is viewport-based raster tiles with cache control, similar to terrain tiles.
+
+## Goals
+
+- Render a high-resolution Earth imagery basemap that visually matches the high-precision coastline and country boundary layer.
+- Load imagery by visible bbox / tile key instead of loading a whole-world giant texture.
+- Keep the current global texture only as a low-zoom background, not as the source of truth for coastlines at inspection zoom.
+- Let imagery failures degrade only the imagery layer; high-precision borders and hover must continue working.
+- Keep generated imagery cache out of Git.
+
+## Data Sources
+
+Candidate sources, in recommended order:
+
+- NASA GIBS / Blue Marble / VIIRS style imagery for permissive global coverage and stable tile service behavior.
+- Sentinel-2 cloudless style public imagery if licensing and tile access are acceptable.
+- A self-hosted raster pyramid generated offline from audited global imagery if third-party online tile terms are unsuitable.
+
+The selected source must document:
+
+- license / attribution
+- max zoom and native resolution
+- tile matrix / projection
+- cache policy
+- whether commercial or public deployment is allowed
+
+## Architecture
+
+```text
+global low-zoom texture
+→ visible Earth bbox from camera raycast
+→ Web Mercator tile keys by zoom
+→ raster tile fetch/cache
+→ project tile image patches onto Earth surface
+→ high-precision coastline / border layer remains above imagery
+```
+
+Implementation should mirror the existing terrain tile discipline:
+
+- dedupe in-flight requests
+- LRU cache for decoded images / textures
+- debounce camera movement
+- cancel or ignore stale viewport requests
+- cap max tiles per frame / per view
+- expose loading/error diagnostics
+
+## Rendering Rules
+
+- The high-resolution imagery layer is visual only. It must not define country hover, coastline, or border geometry.
+- The high-precision coastline remains the visual alignment reference.
+- The border layer render order stays above the basemap imagery.
+- Low zoom may use the current global texture for speed.
+- Mid/high zoom overlays imagery tiles only for the visible region plus a small prefetch ring.
+- Do not draw decorative gradients or fake coastlines to hide mismatch.
+
+## Frontend Work
+
+- Add a new `basemap-imagery.js` module instead of expanding `country-boundaries.js`.
+- Add config in `constants.js`:
+  - source URL template
+  - attribution
+  - min/max zoom
+  - tile cache limit
+  - debounce interval
+  - opacity
+  - enable/disable setting
+- Add Earth settings control:
+  - `高清底图`: off / auto / on
+  - default `auto`
+- Add debug counters for:
+  - active tile count
+  - cached tile count
+  - failed tile count
+  - current imagery zoom
+
+## Backend / Ops Work
+
+- If using a third-party tile service directly, document attribution and rate-limit behavior.
+- If proxying tiles, add backend cache with request coalescing and timeout limits.
+- If self-hosting, add an offline builder that writes ignored tile artifacts under a dedicated data directory.
+- Update Nginx static serving if self-hosted raster tiles are used.
+
+## Performance Budget
+
+- Desktop target: keep visible imagery tiles under a configurable cap, initially 64.
+- Mobile target: lower max zoom and tile cap by default.
+- Decode and upload textures incrementally; avoid blocking Earth startup on high-resolution imagery.
+- First Earth paint must still use the existing lightweight global texture.
+
+## Verification
+
+- Compare high-precision coastline against imagery in coastal areas such as southeast China, Taiwan, Hainan, the Korean peninsula, Japan, and island chains in the South China Sea.
+- Verify zooming / panning does not create visible tile thrash or long blank periods.
+- Verify failed imagery requests do not hide borders or break hover.
+- Verify memory stabilizes after repeated pan/zoom due to LRU eviction.
+- Run `/home/ray/.bun/bin/bun run build`.
+
+## User Operation Steps
+
+After implementation, the user should be able to:
+
+1. Open Earth settings.
+2. Set `高清底图` to `auto` or `on`.
+3. Open Earth and zoom into a coastline.
+4. See imagery tiles refine under the high-precision boundary/coastline layer.
+5. Use diagnostics to confirm which imagery zoom and tile source are active.
+
+## Assumptions
+
+- The existing high-precision vector coastline is the alignment reference.
+- This plan improves visual texture fidelity; it does not replace the boundary data pipeline.
+- A single larger static Earth texture is rejected as the primary solution.
--- a/docs/plans/earth-interactable-layer-plan.md
+++ b/docs/plans/earth-interactable-layer-plan.md
@@ -273,7 +273,7 @@ hover / locked 使用少量 overlay：

 - 算力中心保留现有业务 icon，但接入统一 hover / locked / glow。（已完成）
 - 登陆点曾接入同一套 `Points` 渲染，但 pin 类 SVG 在地球边缘会被深度测试裁切；当前保留专用 `THREE.Sprite`，并使用 canvas 生成黄色扁平球，贴到海缆层级。
- TODO：登陆点暂不迁移到完整 Interactable。后续若要统一交互接口，优先考虑 Sprite-backed adapter，只对齐 `getMarkers()`、`getPointerIntersections()`、`setMarkerState()`、`updateVisualState()` 等外观协议，不强行复用 `THREE.Points`、atlas 和跨图层避让。
+- 登陆点暂不迁移到完整 Interactable。后续若要统一交互接口，优先考虑 Sprite-backed adapter，只对齐 `getMarkers()`、`getPointerIntersections()`、`setMarkerState()`、`updateVisualState()` 等外观协议，不强行复用 `THREE.Points`、atlas 和跨图层避让。
 - 检查图例、搜索和 info-card 是否只依赖业务 payload，而不是依赖渲染对象类型。

 ### Phase 4：形成 Earth 图标层规范
--- a/docs/plans/earth-mobile-center-country-highlight-plan.md
+++ b/docs/plans/earth-mobile-center-country-highlight-plan.md
@@ -37,7 +37,7 @@
 ## Non-goals

 - 不改变桌面端 hover 交互。
- 不替换 `countries-admin0.min.geojson` 数据源。
+- 不引入旧低精度国界兜底；移动端中心国家能力必须复用生产 PMTiles/MVT 国界源。
 - 不新增后端 API。
 - 不把国家面填充做成新的 selected country 面状 shader。
 - 不为移动端增加永久准星 UI，除非后续产品明确需要视觉准星。
@@ -249,4 +249,3 @@ mobileCenterHoverGlowOpacity
 3. 性能保护：加入节流、经纬度阈值和禁用态清理。
 4. 验证：本地构建通过，移动端 viewport 手动检查通过。
 5. 调优：根据截图或真机体验微调阻塞条件和节流阈值。
-
--- a/docs/plans/earth-surface-hover-info-plan.md
+++ b/docs/plans/earth-surface-hover-info-plan.md
@@ -0,0 +1,105 @@
+# Earth Surface Hover Info Plan
+
+Status: Implemented.
+
+## Goal
+
+为 Earth 桌面和 compact 鼠标地表 hover 增加可配置的提示内容，让用户可以选择只看国家信息、只看位置数据，或同时查看国家和经纬海拔。
+
+当前地表 hover 已经具备两类信息：
+
+- 命中国家时显示国家、ISO、大洲，并高亮国界。
+- 未命中国家时显示纬度、经度、海拔。
+
+新方案把这两类信息合并成一个清晰的设置项：`悬停提示`。
+
+## User-facing behavior
+
+设置项放在桌面设置和移动端设置的 `视图` 区，使用分段控件：
+
+- `国家`：陆地命中国家时显示国家名、ISO、大洲；太平洋等海洋区域不显示地表 tooltip。
+- `位置`：陆地和海洋都显示纬度、经度、海拔；不触发国家 tooltip 和国家边界 hover 高亮。
+- `完整`：陆地命中国家时显示国家信息和纬度、经度、海拔；海洋区域显示纬度、经度、海拔。
+
+默认值为 `完整`，因为它保留现有国家识别价值，同时满足 hover 时查看经纬海拔的需求。
+
+海洋在 `国家` 模式下保持沉默，而不是显示大洋名称。原因是当前项目没有海域/大洋边界数据源；用经纬度粗判太平洋、大西洋等范围容易产生误导。如果用户需要海洋位置，使用 `位置` 或 `完整`。
+
+## Implementation plan
+
+### Settings state
+
+在 `frontend/public/earth/js/constants.js` 增加：
+
+```js
+export const SURFACE_HOVER_INFO_MODES = {
+  COUNTRY: "country",
+  POSITION: "position",
+  FULL: "full",
+};
+
+export const DEFAULT_SURFACE_HOVER_INFO_MODE =
+  SURFACE_HOVER_INFO_MODES.FULL;
+```
+
+在 `frontend/public/earth/js/controls.js`：
+
+- 将 `EARTH_SETTINGS_VERSION` 从 `10` 升到 `11`。
+- 在 shared settings 中新增 `surfaceHoverInfoMode`。
+- 新增导出：
+  - `getSurfaceHoverInfoMode()`
+  - `setSurfaceHoverInfoMode(mode, { persist, suppressStatus })`
+- normalize 时只接受 `country | position | full`，否则回退到 `full`。
+- reset settings 后恢复为 `full`。
+
+### Settings UI
+
+在桌面设置 `视图` 区和移动端设置 `视图` 区加入同一组按钮：
+
+```html
+<button data-surface-hover-info-mode="country">国家</button>
+<button data-surface-hover-info-mode="position">位置</button>
+<button data-surface-hover-info-mode="full">完整</button>
+```
+
+控件同步规则沿用现有卫星显示风格和巡航模块的模式：
+
+- 当前模式按钮添加 `is-active`。
+- 当前模式按钮设置 `aria-pressed="true"`。
+- 切换后保存到 Earth settings localStorage。
+
+### Hover tooltip logic
+
+在 `frontend/public/earth/js/main.js` 的地表 hover 分支中读取 `getSurfaceHoverInfoMode()`，统一构造 tooltip。
+
+行为规则：
+
+- 如果没有命中地球：清除国家 hover 并隐藏 tooltip。
+- `position`：
+  - 调用 `clearCountryBoundaryHover()`。
+  - 显示 `纬度 / 经度 / 海拔`。
+- `country`：
+  - 仅当国界图层开启并命中国家时显示国家 tooltip 和国界 hover。
+  - 海洋、国界图层关闭、未加载国界数据时隐藏地表 tooltip。
+- `full`：
+  - 国界图层开启且命中国家时显示国家信息加位置信息。
+  - 未命中国家或国界图层关闭时显示位置信息。
+
+海拔继续使用 `sampleElevationAt(lat, lon)`。暂无采样时显示 `—`，不因 hover 主动加载地形瓦片。
+
+## Acceptance criteria
+
+1. `完整` 模式下，hover 陆地显示国家信息和经纬海拔，hover 海洋显示经纬海拔。
+2. `国家` 模式下，hover 陆地显示国家信息，hover 太平洋等海洋不显示地表 tooltip。
+3. `位置` 模式下，hover 陆地和海洋都显示经纬海拔，国家边界不高亮。
+4. 关闭国界图层后，`完整` 模式回退为只显示位置。
+5. 船只、BGP、算力中心、海缆等对象 hover tooltip 优先级不变。
+6. 移动端中心国家高亮不受这个鼠标 hover 设置影响。
+7. 设置刷新后保持，重置后恢复为 `完整`。
+
+## Verification
+
+- 在 `frontend` 下运行 `/home/ray/.bun/bin/bun run build`。
+- 手动验证三种模式的陆地和海洋 hover 行为。
+- 验证设置持久化和重置。
+- 验证对象 hover 仍优先于地表 hover。
--- a/docs/plans/earth-vessel-ais-aggregation-plan.md
+++ b/docs/plans/earth-vessel-ais-aggregation-plan.md
@@ -176,7 +176,7 @@ freshness:

 ## 聚合接口

-状态更新：开发期已直接切换到新船只快照接口。旧 `/api/v1/visualization/geo/vessels` 不再兼容返回数据，而是返回 `410 Gone`；新的 Earth 船只首屏应调用 `/api/v1/vessels/snapshot`，实时更新走 `/ws` 的 `vessels` 订阅。
+状态更新：开发期已直接切换到新船只快照接口。旧 `/api/v1/visualization/geo/vessels` 路由已移除；新的 Earth 船只首屏应调用 `/api/v1/vessels/snapshot`，实时更新走 `/ws` 的 `vessels` 订阅。

 现有展示接口应逐步改为消费聚合服务，而不是自己直接拼 `VesselPosition + VesselStatic`。

@@ -346,11 +346,11 @@ VesselFinder 等服务里的船只图片不属于 AIS 实时数据本身。图
 5. 船名标准化会读取 AISStream `MetaData.ShipName`；船型展示会从 `vessel_type_name` 和 AIS 数字 `vessel_type` 共同归一化，保证 marker 颜色、详情卡、hover 和搜索结果一致。
 6. 当前实现已转向 `/api/v1/vessels/snapshot`：必须带 bbox / zoom，默认 `limit=1000`，最大 `limit=5000`，不再支持旧 `/geo/vessels` 全量返回。

-### v3.1 — 聚合完整性修复（已被新快照接口取代）
+### v3.1 — 聚合完整性修复（已被受控 fallback 取代）

-原目标是先保证“所有已采集到的船都能显示”，BarentsWatch 不因为接入 AISStream 而被 raw observation 聚合结果遮蔽。开发期产品尚未上线后，决策调整为直接淘汰 legacy 船只表兜底：船只快照只读取 `ais_raw_observations` 聚合结果，旧 `vessel_position + vessel_static` 不再合并进 `/api/v1/vessels/snapshot`。
+原目标是先保证“所有已采集到的船都能显示”，BarentsWatch 不因为接入 AISStream 而被 raw observation 聚合结果遮蔽。当前实现已经移除旧 `/geo/vessels` 路由，船只入口统一为 `/api/v1/vessels/snapshot`。snapshot 优先读取 `ais_raw_observations` 聚合结果；当当前 raw 窗口为空时，才受控回退到 `vessel_position + vessel_static` 最新点，并通过 `diagnostics.legacy_fallback_used` 标记。

-因此以下 legacy merge 要求作废，保留在文档中只作为历史决策记录：
+因此以下旧 `/geo/vessels` 全量 merge 要求作废，保留在文档中只作为历史决策记录：

 1. `/geo/vessels` 必须合并 raw observation 聚合结果和 legacy latest position 结果。
 2. raw 与 legacy 同一 MMSI 同时存在时只显示一艘，优先使用 raw 聚合结果及其 `field_sources` / `selected_reasons`。
@@ -462,8 +462,8 @@ REST collector 的自然状态是 `fetch -> transform -> save -> progress 0..100
 - 明显异常位置不会进入默认展示轨迹，并会留下 `quality_flags`。
 - 同一时间窗口内多来源相近轨迹点只展示一个点。
 - AISStream 重连或回放导致的重复消息不会重复进入聚合结果。
- `/api/v1/vessels/snapshot` 只读取 AIS raw observation 聚合结果；legacy latest position 不再参与船只快照。
- `/api/v1/visualization/geo/vessels` 返回 `410 Gone`，客户端必须迁移到新 snapshot API。
+- `/api/v1/vessels/snapshot` 优先读取 AIS raw observation 聚合结果；当当前 raw 窗口为空时，允许受控 fallback 到 legacy latest position。
+- `/api/v1/visualization/geo/vessels` 路由已移除，客户端必须迁移到新 snapshot API。
 - AISStream 长连接收到新船、位置变化和航向变化后，会通过内部 `/ws` 的 `vessels` channel 推送增量。
 - AISStream streaming 状态不会显示成固定百分比完成进度条，也不会在收到一批消息后误报采集完成。
 - `mmsi`、`imo`、`callsign` 等身份编号在前端不显示千分位符。
--- a/docs/plans/earth-vessel-rendering-performance-plan.md
+++ b/docs/plans/earth-vessel-rendering-performance-plan.md
@@ -216,7 +216,7 @@ hover、locked、dimmed 可通过更新少量 instance attribute 实现，不再

 ### 1. 请求视口范围

-前端请求 `/api/v1/vessels/snapshot` 时必须带上当前视口 `bbox`、`zoom` 和受控 `limit`，减少无关船只。旧 `/api/v1/visualization/geo/vessels` 已下线并返回 `410 Gone`。
+前端请求 `/api/v1/vessels/snapshot` 时必须带上当前视口 `bbox`、`zoom` 和受控 `limit`，减少无关船只。旧 `/api/v1/visualization/geo/vessels` 路由已移除。

 ### 2. 后端排序策略

--- a/docs/plans/earth-vessel-tracking-plan.md
+++ b/docs/plans/earth-vessel-tracking-plan.md
@@ -1,14 +1,14 @@
 # 实时船只监控系统 — 实施计划

-**状态**：规划中
+**状态**：历史计划；实时 AIS 与聚合接口已由 [AIS 多源采集、冲突记录与聚合接口计划](/home/ray/dev/linkong/planet/docs/plans/earth-vessel-ais-aggregation-plan.md) 接管
 **创建日期**：2026-04-27
-**优先数据源**：BarentsWatch AIS（免费但需要 OAuth client credentials）→ AISHub / MarineTraffic（TODO，付费）
+**优先数据源**：BarentsWatch AIS（免费但需要 OAuth client credentials）→ AISStream realtime；AISHub / MarineTraffic 保留为付费备选

 ## 已确认决策

 | 项目 | 决策 |
 |-----|------|
-| 数据源 | BarentsWatch 先行；AISHub / MarineTraffic TODO |
+| 数据源 | BarentsWatch 先行；AISStream realtime 已成为全球实时补充；AISHub / MarineTraffic 保留为付费备选 |
 | 船只规模 | BarentsWatch 阶段全部显示；全球数据接入后按需加船型过滤（默认 Cargo + Tanker + Passenger） |
 | 更新频率 | 准实时：前端 5 分钟轮询，后端 Collector 每分钟拉取写库 |
 | 历史轨迹 | 保留（`vessel_position` 表保留 24h，后期按需扩展） |
@@ -23,9 +23,9 @@
 | 来源类型 | 典型服务 | 覆盖范围 | 成本 | 状态 |
 |---------|---------|---------|------|------|
 | **BarentsWatch AIS API** | live.ais.barentswatch.no | 挪威海域实时 | 免费，需要 AIS API client credentials | **当前使用** |
-| **AISHub** | aishub.net | 全球实时 | 免费/小额 | TODO：付费接入 |
-| **MarineTraffic API** | marinetraffic.com | 全球实时 | $50–$500/月 | TODO：评估 tier |
-| **VesselFinder API** | vesselfinder.com | 全球实时 | $50–$300/月 | TODO：备选 |
+| **AISHub** | aishub.net | 全球实时 | 免费/小额 | 付费备选 |
+| **MarineTraffic API** | marinetraffic.com | 全球实时 | $50–$500/月 | 待评估 tier |
+| **VesselFinder API** | vesselfinder.com | 全球实时 | $50–$300/月 | 备选 |
 | **自建 SDR 接收** | RTL-SDR + AIS-catcher | 仅本地 30–50km | 硬件 $30 | 不考虑 |
 | **NOAA 历史数据** | Marine Cadastre | 美国近海历史 | 免费 | 可用于冷启动 |

@@ -36,16 +36,9 @@
 - 字段：mmsi, lat, lon, sog, cog, heading, nav_status, name, vessel_type, flag
 - 刷新频率：数据约 30–60s 更新一次，可随意轮询

-### TODO：多源 AIS 与实时流接入
+### 多源 AIS 与实时流接入历史

- [ ] 接入 AISStream WebSocket 采集器，作为 BarentsWatch 覆盖不足的实时补充
- [ ] 将 BarentsWatch、AISStream、自定义 `vessel_ais` 映射源统一写入原始观测层
- [ ] 通过聚合接口做去重、字段合并、冲突记录和默认来源选择
- [ ] 开放字段级聚合策略配置，让用户决定不同字段优先信任哪个来源
- [ ] 评估 AISHub 订阅（全球覆盖，约 $30/月），接入全球实时流
- [ ] 评估 MarineTraffic API tier，对比 AISHub 数据质量与成本
- [ ] 实现多数据源适配器，通过 `datasource_config` 切换
- [ ] 真实高频 AIS 稳定接入后，评估将 `vessel_position` 迁移为 TimescaleDB hypertable（保留 Postgres 原生分区作为备选）
+AISStream WebSocket collector、`/api/v1/vessels/snapshot` 和 `/ws` vessels channel 已在后续计划中落地。仍有价值的后续项集中维护在根目录 [TODO](/home/ray/dev/linkong/planet/TODO.md) 的 AIS / Vessels 小节。

 多源 AIS 的详细设计见 [AIS 多源采集、冲突记录与聚合接口计划](/home/ray/dev/linkong/planet/docs/plans/earth-vessel-ais-aggregation-plan.md)。

@@ -120,11 +113,12 @@ CREATE UNIQUE INDEX ON vessel_latest(mmsi);

 #### 1.3 API 端点

-```
-GET /api/v1/visualization/geo/vessels
-    ?bbox=lon_min,lat_min,lon_max,lat_max   # 视口裁剪
+```http
+GET /api/v1/vessels/snapshot
+    ?bbox=lon_min,lat_min,lon_max,lat_max   # 必填，视口裁剪
+    ?zoom=12                                # 必填，当前缩放
    ?type=cargo,tanker,passenger             # 船型过滤
-    ?limit=0                                 # 可选；不传或 0 表示不裁剪数量
+    ?limit=1000                              # 默认 1000，最大 5000
 → GeoJSON FeatureCollection（Point）

 GET /api/v1/visualization/vessels/{mmsi}           # 单船详情
@@ -163,7 +157,7 @@ GeoJSON Feature 格式：
 - 后端 BarentsWatch collector 继续以 HTTP polling 方式采集
 - AISStream 等实时源以独立 WebSocket collector 写入原始观测层
 - 展示接口从聚合服务读取当前船只视图，而不是由单个 collector 决定最终展示值
- 前端默认不再给 `/geo/vessels` 传 `limit=5000`，`VESSEL_CONFIG.maxRenderedMarkers = 0` 表示不做前端数量裁剪；后续如性能不足再引入显式 LOD 上限
+- 旧 `/api/v1/visualization/geo/vessels` 路由已移除，前端必须使用受控 snapshot 接口。
 - marker 颜色、详情卡、hover 和搜索结果必须共享 `vessel_type_display` 船型归一化结果，避免 AIS 数字类型码已驱动颜色但卡片仍显示 `Other`
 - 前端是否升级为 WebSocket delta push 是独立优化，不影响后端采集器可以使用 WebSocket 接上游实时源

@@ -251,7 +245,7 @@ IMO         9811000
 ### Phase 4 — 性能与生产化（2–3 天）

 - `vessel_position` 按天分区，7 天自动清理
- TODO：真实数据量达到百万级/日后，将 `vessel_position` 升级为 TimescaleDB hypertable，配置 retention policy 与压缩策略
+- 真实数据量达到百万级/日后，再评估是否将船只时序数据升级为 TimescaleDB hypertable，并配置 retention policy 与压缩策略
 - GeoJSON endpoint 用 Redis 缓存 15s
 - 若需 bbox 精确查询，引入 PostGIS `geography` + `ST_DWithin`
 - InstancedMesh + frustum culling，目标 5 万船只 60fps
--- a/docs/plans/earthfeed-coordinate-queue-plan.md
+++ b/docs/plans/earthfeed-coordinate-queue-plan.md
@@ -0,0 +1,131 @@
+# EarthFeed 新闻坐标与异步精修计划
+
+## 目标
+
+EarthFeed 的每条新闻都直接携带巡航可用坐标。初始响应使用新闻所属大区的锚点坐标，后台通过消息队列异步推理更精确的目标地址，完成后用实时补丁替换原新闻坐标，实现前端无感更新。
+
+## 返回结构
+
+`GET /api/v1/news/earth-feed` 返回：
+
+```json
+{
+  "generated_at": "2026-05-15T03:16:43Z",
+  "focus": {
+    "lat": null,
+    "lon": null,
+    "region": "global",
+    "label": "全球焦点",
+    "accent": "#d6e6ff"
+  },
+  "sources": [
+    {
+      "id": "bbc-world",
+      "name": "BBC World",
+      "region": "global",
+      "homepage_url": "https://www.bbc.com/news/world"
+    }
+  ],
+  "items": [
+    {
+      "id": "bbc-world:af01519ba7dd",
+      "title": "Flattery and fanfare as Trump welcomed to China - but thorny issues remain",
+      "summary": "The leaders of the world's two superpowers were all smiles...",
+      "url": "https://www.bbc.com/news/articles/cdxpypg9dgeo",
+      "source": "BBC World",
+      "feed_name": "BBC World",
+      "region": "global",
+      "homepage_url": "https://www.bbc.com/news/world",
+      "published_at": "2026-05-14T13:02:13Z",
+      "latitude": 39.9057136,
+      "longitude": 116.3912972,
+      "location_label": "北京市, 中国",
+      "location_source": "headline_location_hint",
+      "verified": true,
+      "location_meta": {
+        "resolution_stage": "headline_location_hint",
+        "ai_attempted": false,
+        "ai_status": "skipped_text_hint",
+        "ai_error": null,
+        "debug_note": "text hint matched 北京市, 中国",
+        "target": {
+          "latitude": 39.9057136,
+          "longitude": 116.3912972,
+          "label": "北京市, 中国",
+          "source": "headline_location_hint",
+          "confidence": 0.78,
+          "country": "中国",
+          "city": "Beijing"
+        },
+        "anchor": {
+          "region": "global",
+          "label": "全球",
+          "latitude": 20.0,
+          "longitude": 0.0
+        }
+      },
+      "is_focus_match": true
+    }
+  ],
+  "errors": [],
+  "stale": false
+}
+```
+
+字段规则：
+
+- `latitude` / `longitude`：前端巡航唯一读取的坐标。
+- `location_label`：当前坐标展示名。
+- `location_source`：`region_anchor`、`headline_location_hint`、`headline_country_hint`、`ai_inferred_target` 等。
+- `verified`：`false` 表示仍是大区锚点；`true` 表示已经由标题规则、国家规则或 AI 得到目标地址。
+- `location_meta`：调试、诊断、AI 状态、目标地址和锚点详情都放这里，不再展开成 `t_*` 主字段。
+
+## 后台队列
+
+当前使用 Redis Streams：
+
+- stream：`earth_news:target_location:jobs`
+- consumer group：`earth_news_target_location`
+- result cache：`earth_news:target_location:result:{item_id}`
+- dedupe key：`earth_news:target_location:queued:{item_id}`
+
+请求流程：
+
+1. RSS 拉取并排序。
+2. 每条新闻先生成大区锚点坐标，`verified=false`。
+3. 若 Redis 已有该新闻的精修结果，则合并结果返回。
+4. 若没有精修结果，则把新闻 job 入队，接口立即返回。
+
+Worker 流程：
+
+1. 从队列消费新闻 job。
+2. 先跑标题/国家规则，再视情况调用 AI。
+3. 写入 result cache。
+4. 广播 WebSocket 补丁：
+
+```json
+{
+  "type": "data_frame",
+  "channel": "earth_news",
+  "timestamp": "2026-05-15T03:17:00Z",
+  "payload": {
+    "item_id": "bbc-world:af01519ba7dd",
+    "patch": {
+      "latitude": 39.9057136,
+      "longitude": 116.3912972,
+      "location_label": "北京市, 中国",
+      "location_source": "headline_location_hint",
+      "verified": true,
+      "location_meta": {}
+    }
+  }
+}
+```
+
+## 可迁移性
+
+业务代码只调用队列接口，不直接依赖 Redis Streams 细节。以后迁移 Kafka 时新增 Kafka adapter，保持 job payload、result patch 和 worker 推理逻辑不变。
+
+## 前端规则
+
+新闻面板和巡航都只读取新闻项内的 `latitude` / `longitude`。实时补丁到达后按 `item_id` 合并到现有 `payload.items`，重新渲染并触发 `earth:news-payload-updated`，巡航下一轮自然使用精修坐标。
--- a/docs/plans/production-delivery-cicd-stabilization-plan.md
+++ b/docs/plans/production-delivery-cicd-stabilization-plan.md
@@ -0,0 +1,55 @@
+# Planet 正式交付与 CI/CD 稳定化计划
+
+## Summary
+
+- CI/CD 平台采用 Gitea Actions，由自托管 `act_runner` 执行。
+- 正式交付目标采用 Kubernetes，端口、健康检查、重启和滚动发布交给 Service、Ingress、readiness/liveness probe。
+- Vite 继续保留，但只作为开发服务器和生产构建工具；生产运行 nginx 托管 `vite build` 产物。
+- 不新增 Webpack 双构建链。Electron 仅在离线桌面交付成为明确目标后再评估。
+
+## Key Changes
+
+- 生产镜像：
+  - frontend 多阶段构建，`bun install` + `bun run build`，最终 nginx 托管 `dist`。
+  - backend/aiprovider 移除 `--reload`，加入容器健康检查。
+  - 镜像标签使用 `<registry>/<namespace>/<service>:<git-sha>`，发布 tag 额外推 `vX.Y.Z`。
+- Kubernetes：
+  - 新增 Helm chart：`deploy/helm/planet`。
+  - frontend 暴露 Ingress；backend/aiprovider 默认 ClusterIP。
+  - PostgreSQL/Redis 默认外部依赖，`values.single-node.yaml` 提供演示/测试内置依赖。
+- Gitea Actions：
+  - `ci.yaml`：后端测试、前端构建、Docker build smoke、Helm render。
+  - `release.yaml`：构建并推送三类镜像。
+  - `deploy-staging.yaml`：部署 staging、等待 rollout、执行 smoke tests。
+- 开发脚本边界：
+  - `planet.sh` 保留为本地开发便利脚本。
+  - CI/CD 与正式部署不调用 `planet.sh start`。
+
+## Vite / Webpack / Electron Decision
+
+中肯结论：不要因为“企业生产环境”这件事去做 Webpack 版本；继续用 Vite，但把“开发服务器”和“生产构建/部署”分清楚。Electron 也不要现在做，除非正式版目标明确是离线桌面软件。
+
+Vite 可以用于生产构建。生产环境运行的是 `vite build` 产出的静态资源，不是 Vite dev server。当前项目已经使用 React + Vite + Bun、`import.meta.env`、`public/earth` 静态资产路径和大量 Three.js/ES module 资源引用。维护 Webpack 双构建链会显著增加路径、资源、环境变量和回归测试成本。
+
+如果未来客户环境确实要求更接近 Webpack 生态，优先做 Rsbuild/Rspack 技术 spike，而不是直接维护 Webpack 并行构建。Electron 适合离线运行、本地硬件/文件访问、系统托盘、自动更新和安装包分发；但 Planet 目前还包含 backend、database、Redis、AI Provider、Motion Agent 等服务编排，桌面壳不能解决正式交付的核心问题。
+
+## Test Plan
+
+- CI gates：
+  - `uv sync --group dev`
+  - `uv run pytest backend/tests/test_api.py backend/tests/test_realtime_sources.py -q`
+  - `cd frontend && bun install --frozen-lockfile && bun run build`
+  - Docker build frontend/backend/aiprovider
+  - `helm lint deploy/helm/planet`
+  - `helm template planet-staging deploy/helm/planet -f deploy/helm/planet/values.single-node.yaml`
+- Staging deployment:
+  - `helm upgrade --install planet-staging deploy/helm/planet --namespace planet-staging`
+  - 等待 frontend/backend/aiprovider rollout。
+  - smoke test frontend `/`、frontend `/health`、backend `/health`、aiprovider `/health`。
+
+## Acceptance Criteria
+
+- main/dev 提交能通过 CI。
+- 发布 workflow 能生成可追踪镜像。
+- staging 可从零部署并完成滚动升级。
+- 正式部署不依赖本机端口清理，也不运行 Vite dev server。
--- a/docs/technical/en/agents-aiprovider.md
+++ b/docs/technical/en/agents-aiprovider.md
@@ -95,9 +95,24 @@ The AI settings page uses:
 - `POST /api/v1/settings/integrations/ai-provider/connect`
 - `GET /api/v1/settings/integrations/ai-provider/secrets`
 - `GET /api/v1/settings/integrations/ai-provider/presets`
+- `GET /api/v1/settings/ai-prompts`
+- `PUT /api/v1/settings/ai-prompts/{task_key}`
+- `POST /api/v1/settings/ai-prompts/{task_key}/reset`

 These endpoints require an authenticated user. The `secrets` endpoint is only used when the settings page reveals a key or token; hiding the field restores the masked preview.

+The `ai-prompts` endpoints back the Prompts tab in AI settings. Shipped defaults come from versioned backend resources, while business code references stable task keys. The API stores only operator overrides. Resetting a prompt removes the override and falls back to the current shipped default.
+
+### Prompt Boundary
+
+`aiprovider` is a pure model adapter and does not inject a global business system prompt. News localization, alert briefing, BGP briefing, location factcheck, datasource mapping, and credential guide generation each resolve their own effective prompt by task key. Alert-analysis system prompts are only sent by alert-related tasks and do not leak into other LLM calls.
+
+### Agent And Tool Boundary
+
+Agent workflows belong in the `backend`, not in `aiprovider`. Future Earth LLM commands, situational awareness, multi-role simulation, WebSearch, database queries, evidence storage, and configuration proposal application should be orchestrated by the backend Agent Runtime. `aiprovider` should receive model-ready requests from the backend and return normalized model responses.
+
+If a provider supports native tool calling, `aiprovider` may pass through protocol fields and normalize response blocks, but tool whitelists, argument validation, permission policy, run records, and write approvals must stay in the backend. When a provider does not support native tools, the backend uses JSON tool-call fallback; business tools should not move into `aiprovider` for a provider-specific shortcut.
+
 ### AI provider internal API

 Internal-only endpoints:
@@ -287,7 +302,6 @@ SERVICE_VERSION=0.1.0
 AI_PROVIDER_SERVICE_TOKEN=change_me
 AI_TIMEOUT_SECONDS=60
 AI_HTTP_RETRY_ATTEMPTS=2
-AI_ANALYSIS_SYSTEM_PROMPT=你是态势感知分析助手。请基于输入的上下文、观测与约束，输出结构化、克制、可执行的分析。
 ```

 Optional provider-specific keys:
--- a/docs/technical/en/backend-collectors.md
+++ b/docs/technical/en/backend-collectors.md
@@ -89,6 +89,8 @@ async def run(self, db):

 AIS vessel collectors use a different persistence path from regular `CollectedData` collectors. BarentsWatch, AISStream, and custom `vessel_ais` sources write into the AIS raw observation layer first, then the aggregation service merges those observations into the GeoJSON and detail payloads used by the Earth vessel layer. This preserves source, transport, field conflicts, and observation time instead of letting one realtime source overwrite the final display table.

+Earth boundaries are no longer data collectors. They are Earth static rendering assets: the Earth Assets settings panel owns source configuration, and `/api/v1/earth/boundaries/*` builds `frontend/public/earth/data/boundaries/earth-boundaries-china-pov-v1.pmtiles`. When no high-precision PMTiles artifact is available locally, the frontend uses the bundled low-precision GeoJSON fallback and does not write boundary records to `CollectedData`.
+
 TOP500 and Epoch AI compute sources do not always provide usable coordinates. The unified Earth compute-center endpoint uses only valid source-provided coordinates or `compute_center_locations` dimension-table coordinates during the main map startup path; records without coordinates are returned as `unresolved` instead of being rendered from a local registry, country centroid, or guessed city. When users manually collect candidates, the backend queries ROR and Nominatim/OpenStreetMap from source fields; accepted candidates are saved into `compute_center_locations` and rendered from that table on the next layer refresh.

 ## IV. Data Format (stored in CollectedData table)
@@ -219,7 +221,8 @@ backend/app/services/collectors/
 ├── peeringdb.py         # PeeringDB collector
 ├── telegeraphy.py       # TeleGeography submarine cable collector
 ├── vessel_ais.py        # BarentsWatch AIS vessel collector
-└── aisstream.py         # AISStream WebSocket vessel collector
+├── aisstream.py         # AISStream WebSocket vessel collector
+└── earth_boundaries.py  # Earth boundary source verification and static tile artifact collector

 backend/app/services/
 ├── custom_datasource_runtime.py      # Custom REST / WebSocket mapping runtime
@@ -297,7 +300,18 @@ State semantics:
 - `reconnecting`: upstream or network interruption; the collector records `AISSourceHealth` and waits before reconnecting.
 - `stopped` / `cancelled`: stopped by a test limit or user action.

-AISStream connectivity validation reads the saved collector configuration, environment variables, and `AISSTREAM_API_KEY` in `~/.zshrc` through `datasource_connectivity.py`. For actual collection, the most reliable path is saving the API key in `Settings -> Collector Settings -> AISStream Vessels`; if the key only lives in `~/.zshrc`, confirm that the backend process inherited it.
+AISStream connectivity validation reads the saved collector configuration, environment variables, and `AISSTREAM_API_KEY` in `~/.zshrc` through `datasource_connectivity.py`. For actual collection, the most reliable path is saving the API key in `Collection Management -> Collectors -> AISStream Vessels`; if the key only lives in `~/.zshrc`, confirm that the backend process inherited it.
+
+The console manages AISStream from `/datasources -> Realtime Streams`, not from the normal finite collection progress bar. The realtime stream API aggregates runtime state, health, configuration preview, and raw observation counters:
+
+```http
+GET /api/v1/realtime-sources
+POST /api/v1/realtime-sources/{source}/start
+POST /api/v1/realtime-sources/{source}/stop
+POST /api/v1/realtime-sources/{source}/restart
+```
+
+`aisstream_vessels` and custom `source_type=websocket` sources appear in that API. They do not participate in one-click collection percentages; the UI interprets them as long-lived services with message counters, lag, last success, and last error.

 ### AIS Raw Observations And Aggregation

@@ -318,7 +332,7 @@ GET /api/v1/visualization/vessels/{mmsi}/track
 GET /api/v1/visualization/vessels/{mmsi}/conflicts
 ```

-`/api/v1/vessels/snapshot` requires `bbox` and `zoom`, defaults to `limit=1000`, and caps `limit` at `5000`. It reads only aggregated `ais_raw_observations`; it no longer merges legacy `vessel_position` / `vessel_static` rows. The old `/api/v1/visualization/geo/vessels` endpoint has been removed and returns `410 Gone`.
+`/api/v1/vessels/snapshot` requires `bbox` and `zoom`, defaults to `limit=1000`, and caps `limit` at `5000`. It prefers aggregated `ais_raw_observations`; when the current raw window is empty, it can fall back to the latest legacy `vessel_position` / `vessel_static` rows and marks that path with `diagnostics.legacy_fallback_used`. The old `/api/v1/visualization/geo/vessels` route has been removed.

 Realtime deltas are sent through the `/ws` `vessels` channel. Clients must subscribe with the current viewport:

@@ -354,9 +368,9 @@ GET /api/v1/layers/bgp/collectors?bbox=lon_min,lat_min,lon_max,lat_max&zoom=12&l

 `/api/v1/data-products/*` is for aggregate panels and keeps a global statistics scope independent of the map bbox. `/api/v1/layers/*` is for map rendering, requires `bbox` and `zoom`, defaults to `limit=1000`, and caps `limit` at `5000`; low zoom falls back to a smaller response cap and reports `degraded`, `truncated`, `limit_clamped`, and `stats_scope=viewport` in `diagnostics`. Non-vessel layers currently reuse the existing GeoJSON converters before the guard layer; future product-specific queries can push bbox filtering deeper.

-## X. Collector Settings And Connectivity Validation
+## X. Collectors And Connectivity Validation

-The console "Collector Settings" page owns endpoint, headers, timeouts, retries, and credentials for all built-in collectors. Connectivity is derived by the backend checksum rather than by frontend button styling:
+The console "Collectors" page owns endpoint, headers, timeouts, retries, and credentials for all built-in collectors. Connectivity is derived by the backend checksum rather than by frontend button styling:

 - endpoint
 - auth type
@@ -377,7 +391,7 @@ POST /api/v1/settings/credential-guides/{provider}/generate
 POST /api/v1/settings/credential-guides/{provider}/reset
 ```

-See [Collector Settings and Connectivity Validation](/home/ray/dev/linkong/planet/docs/technical/en/datasource-collector-settings-connectivity.md) for the full flow.
+See [Collectors and Connectivity Validation](/home/ray/dev/linkong/planet/docs/technical/en/datasource-collector-settings-connectivity.md) for the full flow.

 ## XI. Data Usage

--- a/docs/technical/en/datasource-collector-settings-connectivity.md
+++ b/docs/technical/en/datasource-collector-settings-connectivity.md
@@ -8,12 +8,12 @@ The console now separates the "data source catalog" from "collector configuratio
  - Lists all data sources, including built-in and custom sources.
  - Clicking a name only opens an information drawer.
  - Focuses on status, manual collection, and running collection tasks.
- `/settings?tab=collector_credentials`
-  - Displays as "Collector Settings".
+- `/collection-management?tab=collector_credentials`
+  - Displays as "Collectors".
  - Owns endpoint, headers, base parameters, and credentials.
  - Every collector exposes a connection button for health checks.

-This reduces first-use confusion: API endpoints, headers, credentials, and custom source configuration all belong to collector settings instead of being scattered across the data source list and system settings.
+This reduces first-use confusion: API endpoints, headers, credentials, and custom source configuration all belong to Collectors instead of being scattered across the data source list and system settings.

 ## User-Facing Rules

@@ -53,7 +53,7 @@ Current behavior:

 `data-source-bulk-toolbar__running-pill` is the styling entry point for the "Collecting" pill. It is aligned with other status tags, while hover treatment, arrow affordance, and blue outline indicate interactivity.

-### Collector Settings
+### Collectors

 File:

@@ -61,7 +61,7 @@ File:

 Current behavior:

- The `collector_credentials` tab is displayed as "Collector Settings".
+- The `collector_credentials` tab is displayed as "Collectors" under `/collection-management`.
 - A select lists built-in collectors and supports maintaining custom supplemental sources that merge into built-in data.
 - The only button beside the select is a plug icon for health checks.
 - Status tags below the select show:
@@ -288,7 +288,7 @@ Normalization:

 Connectivity validation reads saved configuration, environment variables, and `AISSTREAM_API_KEY` from `~/.zshrc`. For actual collection, prefer saving the API key in collector settings. If the key only lives in `~/.zshrc`, confirm that the backend process inherited it; otherwise validation may pass while the collector runtime cannot read the key.

-Connectivity validation and actual collection are separate actions. A banner such as `AISStream credentials configured, WebSocket endpoint format valid` only means the saved settings can be used for a connection attempt; runtime status may still be `disconnected`. Global AIS data is written locally only while the `aisstream_vessels` collector is `streaming` / `connected` and its message count plus `last_seen_at` keep advancing.
+Connectivity validation and actual collection are separate actions. A banner such as `AISStream credentials configured, WebSocket endpoint format valid` only means the saved settings can be used for a connection attempt; runtime status may still be `disconnected`. Global AIS data is written locally only while `aisstream_vessels` is `streaming` / `connected` and its realtime stream counters plus `last_seen_at` keep advancing. Start, stop, reconnect, health, and counters are exposed through `/datasources -> Realtime Streams` and `/api/v1/realtime-sources`; AISStream is not counted in normal one-click collection progress.

 The new vessel list entry point is no longer the legacy `/api/v1/visualization/geo/vessels` route. Earth initial state should call:

@@ -296,7 +296,7 @@ The new vessel list entry point is no longer the legacy `/api/v1/visualization/g
 GET /api/v1/vessels/snapshot?bbox=lon_min,lat_min,lon_max,lat_max&zoom=12&limit=1000
 ```

-That endpoint reads local aggregated `ais_raw_observations` only. Realtime updates use the `/ws` `vessels` channel; subscriptions must include `bbox`, `zoom`, and `limit`. The server filters updates per connection and merges collector broadcasts every second, keeping only the latest position per MMSI.
+That endpoint prefers local aggregated `ais_raw_observations`; when the current raw window is empty, it can fall back to the latest legacy `vessel_position` / `vessel_static` rows and exposes that through `diagnostics.legacy_fallback_used`. Realtime updates use the `/ws` `vessels` channel; subscriptions must include `bbox`, `zoom`, and `limit`. The server filters updates per connection and merges collector broadcasts every second, keeping only the latest position per MMSI.

 ## Custom REST / WebSocket Mapping Runtime

@@ -307,6 +307,8 @@ Files:

 Custom sources are supplemental inputs for existing target schemas, not isolated data islands. The most complete target today is `vessel_ais`: a custom REST or WebSocket source is mapped deterministically, written into AIS raw observations, and then pushed to Earth through the `vessels` WebSocket channel.

+Earth high-precision boundaries no longer use custom-source target schemas. Boundaries are Earth static assets: the Earth Assets settings panel saves local source configuration and triggers PMTiles builds without writing records to `CollectedData`.
+
 ### Configuration Semantics

 Important fields:
@@ -316,7 +318,7 @@ Important fields:
 - `auth_type`: `none`, `bearer`, `api_key`, or `basic`.
 - `headers`: static request headers.
 - `auth_config`: token, API key, or basic username/password; API keys can be sent by header or query.
- `config.target_schema`: for example `vessel_ais`.
+- `config.target_schema`: for example `vessel_ais`, `geo_points`, or `generic_records`.
 - `config.delivery_mode`: REST defaults to `polling`; WebSocket defaults to `realtime_stream`.
 - `config.merge_target_source`: records which built-in source this custom source supplements, such as `barentswatch_vessels`.

--- a/docs/technical/en/earth-bgp-context.md
+++ b/docs/technical/en/earth-bgp-context.md
@@ -157,6 +157,7 @@ Current design:
   - symbol-driven event cores
   - outward ring pulses
   - reduced diffuse glow compared with older Earth builds
+   - event region halos, collector coverage halos, and radar pulses derive a lighter tint from their own icon color instead of using a fixed teal; red high-severity events, orange active collectors, and blue idle collectors keep their hue family while broad halos stay softer than the icon
 5. The right-side stats now show:
   - BGP events
   - collector count
--- a/docs/technical/en/earth-frontend-context.md
+++ b/docs/technical/en/earth-frontend-context.md
@@ -71,6 +71,10 @@ Responsibilities:

 This is currently the most critical UI control entry point for the Earth frontend.

+Earth settings are now grouped by `data-settings-tab` and `data-settings-tab-panel`. Desktop and mobile share the same category semantics: Runtime, Display, Panels, Motion, Shortcuts, and System. When adding a setting, first choose its category, then add the DOM, persistence field, and restore logic; do not keep growing one long undifferentiated panel.
+
+Shortcut configuration is a device-local preference owned by `controls.js`: read, capture, enable/disable, and reset all stay in the Earth frontend. It should not be written to backend user settings and should not affect other browsers. New shortcuts must provide a default key, display label, disabled/enabled state, and reset path instead of being hard-coded only in a keydown handler.
+
 ### 4. UI and Status Messages

 - [ui.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/ui.js)
@@ -122,6 +126,9 @@ Responsibilities:
 - Globe sphere, cloud layer, atmosphere
 - Real terrain mesh
 - Terrain tile fetch, decode, displacement, and shading
+- Whole-globe land/ocean and border base overlays
+
+The Earth surface is a stack of near-concentric shells, not a single mesh. The base sphere and HD texture overlay in `earth.js`, plus the land/ocean base in `country-boundaries.js`, need explicit radius separation. At far zoom, GPU depth precision drops; neighboring shells that are too close can z-fight and show black flicker blocks or snow. The current stable spacing is `landAltitudeOffset = 0.32` and `textureOverlayAltitudeOffset = 0.48`. When adding or adjusting whole-globe surface overlays, update [Earth Render Layer Order](/home/ray/dev/linkong/planet/docs/technical/en/earth-render-layer-order.md) and verify at 50% zoom.

 ### 7. Layer Modules

@@ -142,6 +149,8 @@ Each module is responsible for its own:

 `tv.js` owns the live / aggregation-news tabs inside `media-panel`. Toolbar open and tab-switch actions write back through `earth:tv-visibility-change` and `earth:tv-tab-change`: panel visibility remains viewport-scoped at `views.<scope>.panelVisibility.media-panel`, while the active tab is stored at `shared.mediaPanelActiveTab`. Refreshing the page therefore restores the user's last live/news state. Temporary hides from `closeTransientMobileOverlays()` carry `persist:false` and do not overwrite the preference.

+`brand.js` manages Earth HUD brand resources. Static assets provide the default brand; runtime overrides come from `/api/v1/earth/brand`, and uploaded images are served from `/earth-brand-assets/...`. The frontend must treat logo/title images and text fallback separately: if an image fails, show the text title; if text fields are empty, rely on backend defaults so the HUD brand area never renders blank. The console Earth Content page owns saving and resetting brand configuration; the Earth frontend only consumes it.
+
 The compute-center layer row has a notification badge for GeoJSON `unresolved` records. The badge means "no trustworthy coordinates, cannot render on the globe"; it is different from the `?` marker drawn on already positioned but unconfirmed compute centers. Clicking the badge opens a fixed info card beside the layer panel. Row-level `采集` fetches candidates only. Header-level `一键采用` processes the queue top-to-bottom, saves the highest-confidence valid candidate, removes successful rows, renumbers the list, and dispatches `earth:compute-center-unresolved-count-change` so the badge updates immediately. When the batch ends, `earth:compute-center-location-saved` refreshes the real layer.

 Location candidate state in the details card is cached in [info-card.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/info-card.js) by `entityType:entityId`. If the user closes the details card or unresolved queue and reopens the same compute center / BGP collector, previously collected candidates and status text are restored. Header-level `一键采用` prefers cached candidates, avoiding repeated online geocoding or LLM factcheck calls. After a location is saved, that entity's candidate list is cleared to a "refreshing layer" status so stale candidates do not keep misleading the user.
@@ -154,7 +163,7 @@ The `earth:compute-center-location-saved` reconciliation pipeline is deliberatel

 The vessel layer now uses `/api/v1/vessels/snapshot` for the initial viewport snapshot and the `/ws` `vessels` channel for realtime deltas. Snapshot requests must include `bbox`, `zoom`, and a bounded `limit`; the backend defaults to `limit=1000` and caps it at `5000`. WebSocket subscriptions must include the same viewport fields so the server can filter updates per connection.

-The legacy `/api/v1/visualization/geo/vessels` endpoint has been removed and returns `410 Gone`. Frontend code should fetch a snapshot for the current viewport when the layer opens, then subscribe to `vessels` deltas. After map pan or zoom, reload the snapshot and send a fresh vessels subscription. The backend no longer merges legacy `vessel_position` / `vessel_static` rows into vessel snapshots, so the frontend must not depend on old BarentsWatch-only fallback rows.
+The legacy `/api/v1/visualization/geo/vessels` route has been removed. Frontend code should fetch a snapshot for the current viewport when the layer opens, then subscribe to `vessels` deltas. After map pan or zoom, reload the snapshot and send a fresh vessels subscription. The backend only falls back to legacy `vessel_position` / `vessel_static` rows when the current raw window is empty; frontend code can detect that state through `diagnostics.legacy_fallback_used`.

 The new layer API family is `/api/v1/layers/*`, which separates map rendering payloads from aggregate panel statistics. Layer requests must include `bbox`, `zoom`, and a bounded `limit`; responses include `visible_count`, `returned_count`, and `diagnostics`, where `degraded`, `truncated`, and `limit_clamped` are the frontend signals for fallback UI. Right-side aggregate panels should not sum the layer response. They should read `/api/v1/data-products` or `/api/v1/data-products/{product_id}/status`, because those statistics stay global and do not change with the viewport.

@@ -205,7 +214,15 @@ Terrain should not block startup when it is not the restored visible layer. Afte

 Earth settings are stored in `localStorage`. The key is typically a namespaced string defined in `constants.js`. `controls.js` handles read, write, and reset.

-Settings that affect visual layers (terrain opacity, day/night mode, satellite display style, etc.) are read during initialization and applied immediately.
+Settings that affect visual layers and surface interaction (terrain opacity, day/night mode, satellite display style, satellite idle breathing, real satellite altitude, track display, hover tooltip mode, etc.) are read during initialization and applied immediately.
+
+The surface hover tooltip preference is persisted by `controls.js` as `shared.surfaceHoverInfoMode`, while `main.js` composes the actual tooltip in the globe-surface hover branch. `Country` shows country details only when a country polygon is hit and stays silent over ocean; `Position` shows latitude, longitude, and sampled terrain elevation and clears country-boundary hover; `Full` shows country + position on land and position over ocean.
+
+The real satellite altitude preference is persisted by `controls.js`, while the rendering state lives in `satellites.js`. When enabled, the real radius from SGP4 is compressed logarithmically into the current Earth visual radius range. When disabled, satellite dots, trails, and predicted orbits all return to the legacy same-sphere display. Toggling this setting must refresh satellite positions and clear trail buffers so a trail never mixes both height models. `maxRealAltitudeOffset = 25` is a visual cap tuned for the current camera and `earthRadius = 100`: GEO / MEO remain clearly higher than LEO, but the highest orbits stay within about 25% beyond the globe radius so selection targets, red trails, and the globe do not feel disconnected.
+
+SGP4 propagation returns an inertial-frame position, so it must not be drawn directly as Earth-fixed longitude / latitude. `satellites.js` uses `gstime` to convert ECI/TEME positions to ECF, then maps that result into the same Three.js axes as `latLonToVector3()`. Satellite dots and short trails use Earth-fixed coordinates for each sample time, representing the object's current position relative to the globe surface. The locked predicted orbit uses the `gstime` from the lock moment for the whole future orbit, projecting the inertial orbit plane onto the current globe pose; that keeps the line closed and keeps the visual orbit inclination aligned with the details card. Fallback predicted orbits must also use a real RAAN + inclination orbital-plane formula, not treat inclination as a constant latitude.
+
+Boundary precision is stored separately by `country-boundaries.js` under `planet.earth.boundaries.highPrecisionEnabled`. When high precision is off, Earth keeps using the bundled low-precision `countries-admin0.min.geojson` fallback even if high-precision manifest/PMTiles files exist locally. When high precision is on but the artifact is missing, the Earth toolbar settings call `/api/v1/earth/boundaries/build` and poll progress. After success, `reloadCountryBoundaries()` hot-swaps the boundary layer without refreshing the page. Boundary hover is independent from interactable hover: a country polygon remains highlighted whenever the surface coordinate is inside it, while the tooltip can still prioritize a satellite, vessel, BGP marker, or other interactable.

 ## Current Terrain Pipeline

--- a/docs/technical/en/earth-layer-style-reference.md
+++ b/docs/technical/en/earth-layer-style-reference.md
@@ -23,7 +23,7 @@ This document records the material, color, opacity, line width, radius offset, a
 | Earth base specular | `EARTH_MATERIAL_CONFIG.specular` | `0x1a2d45` | `MeshPhongMaterial.specular` |
 | Earth base shininess | `EARTH_MATERIAL_CONFIG.shininess` | `12` | `MeshPhongMaterial.shininess` |
 | Earth base opacity | `EARTH_MATERIAL_CONFIG.opacity` | `1` | `MeshPhongMaterial.opacity` |
-| HD texture radius offset | `EARTH_MATERIAL_CONFIG.textureOverlayAltitudeOffset` | `0.1` | Standalone HD texture sphere radius |
+| HD texture radius offset | `EARTH_MATERIAL_CONFIG.textureOverlayAltitudeOffset` | `0.48` | Standalone HD texture sphere radius; must keep enough depth separation from the land/ocean base and Earth base sphere to avoid far-zoom z-fighting |
 | HD texture opacity | `EARTH_MATERIAL_CONFIG.textureOverlayOpacity` | `0.88` | HD texture `MeshPhongMaterial.opacity` |
 | HD texture renderOrder | `EARTH_MATERIAL_CONFIG.textureOverlayRenderOrder` | `0.96` | `_earthTextureOverlay.renderOrder` |
 | HD texture specular | `EARTH_MATERIAL_CONFIG.textureOverlaySpecular` | `0x05080d` | Reduces specular highlight in direct-light areas to avoid blown-out texture |
@@ -78,11 +78,19 @@ The land/ocean base is an Earth base-map asset and preloads at startup; the "Bor

 | Name | Variable | Current Value | Location / Notes |
 | --- | --- | --- | --- |
-| Country border data path | `COUNTRY_BOUNDARY_CONFIG.dataPath` | `"/earth/data/countries-admin0.min.geojson"` | GeoJSON input |
+| Boundary tile manifest | `COUNTRY_BOUNDARY_CONFIG.tileManifestPath` | `"/earth/data/boundaries/v1/manifest.json"` | High-precision PMTiles manifest; missing manifest uses the low-precision fallback |
+| Boundary tile provider | `COUNTRY_BOUNDARY_CONFIG.tileProvider` | `"auto"` | Prefer high-precision PMTiles/MVT, then fall back to legacy GeoJSON |
+| PMTiles artifact path | `COUNTRY_BOUNDARY_CONFIG.pmtilesPath` | `"/earth/data/boundaries/earth-boundaries-china-pov-v1.pmtiles"` | Production single-file PMTiles/MVT artifact |
+| Low-precision fallback | `COUNTRY_BOUNDARY_CONFIG.legacyFallbackPath` | `"/earth/data/countries-admin0.min.geojson"` | Default land/ocean base and hover data when no high-precision boundary asset has been built locally |
+| MVT layer names | `COUNTRY_BOUNDARY_CONFIG.mvtLayerNames` | `boundary_admin0 / boundary_disputed_internal / coastline / claim_line` | Fixed layer names decoded by the PMTiles provider |
+| Boundary tile base path | `COUNTRY_BOUNDARY_CONFIG.tileBasePath` | `"/earth/data/boundaries/v1/"` | PMTiles manifest base path |
+| Boundary tile zoom thresholds | `COUNTRY_BOUNDARY_CONFIG.tileZoomThresholds` | `1.6 -> z5`, `2.8 -> z6`, `3.4 -> z7`, `4.0 -> z8`, `4.6 -> z9`, `5.2 -> z10` | Production PMTiles zoom selection |
+| Boundary tile cache limit | `COUNTRY_BOUNDARY_CONFIG.tileCacheLimit` | `150` | Frontend LRU cache entries for loaded tile geometries |
+| Boundary tile debounce | `COUNTRY_BOUNDARY_CONFIG.tileDebounceMs` | `180` | View-change debounce before requesting visible tiles |
 | Ocean fill color | local `OCEAN_HEX` | `0x010609` | Land/ocean base canvas background |
 | Land fill color | `COUNTRY_BOUNDARY_CONFIG.landColor` | `0x080f1b` | Land/ocean base canvas land |
 | Land/ocean base opacity | `COUNTRY_BOUNDARY_CONFIG.landOpacity` | `1.0` | `MeshBasicMaterial.opacity` |
-| Land/ocean base radius offset | `COUNTRY_BOUNDARY_CONFIG.landAltitudeOffset` | `0.08` | `country-land-ocean` radius |
+| Land/ocean base radius offset | `COUNTRY_BOUNDARY_CONFIG.landAltitudeOffset` | `0.32` | `country-land-ocean` radius; separated from the Earth base sphere to avoid snow / black block flicker at 50% zoom |
 | Land/ocean base renderOrder | `COUNTRY_BOUNDARY_CONFIG.landRenderOrder` | `0.86` | `country-land-ocean.renderOrder` |
 | Land/ocean mask size | `landMaskWidth / landMaskHeight` | `2048 / 1024` | Canvas / DataTexture size |
 | Country tint color | `COUNTRY_BOUNDARY_CONFIG.tintColor` | `0x0b1830` | Tint when HD texture is off |
@@ -91,16 +99,16 @@ The land/ocean base is an Earth base-map asset and preloads at startup; the "Bor
 | Border line color | `COUNTRY_BOUNDARY_CONFIG.lineColor` | `0x7fc7ff` | Normal border line |
 | Border line opacity | `COUNTRY_BOUNDARY_CONFIG.lineOpacity` | `0.58` | Normal border line opacity |
 | Border dimmed opacity on hover | `COUNTRY_BOUNDARY_CONFIG.dimmedLineOpacity` | `0.18` | Normal border opacity during hover |
-| Border line radius offset | `COUNTRY_BOUNDARY_CONFIG.lineAltitudeOffset` | `0.115` | Normal border line radius; slightly above HD texture `0.10` and below terrain base `0.16` to reduce floating |
+| Border line radius offset | `COUNTRY_BOUNDARY_CONFIG.lineAltitudeOffset` | `0.115` | Normal border line radius; line layers rely on renderOrder and independent geometry, not whole-globe shell depth spacing |
 | Border line renderOrder | `COUNTRY_BOUNDARY_CONFIG.lineRenderOrder` | `2.2` | Normal border line level |
 | Border hover color | `COUNTRY_BOUNDARY_CONFIG.hoverLineColor` | `0xff3b1f` | Neon red-orange |
 | Border hover opacity | `COUNTRY_BOUNDARY_CONFIG.hoverLineOpacity` | `1.0` | Hover line opacity |
-| Border hover radius offset | `COUNTRY_BOUNDARY_CONFIG.hoverAltitudeOffset` | `0.14` | Hover line radius; close to the surface but above normal border lines |
+| Border hover radius offset | `COUNTRY_BOUNDARY_CONFIG.hoverAltitudeOffset` | `0.115` | Hover line radius; matches the normal border geometry to avoid double-edge ghosting during highlight changes |
 | Border hover renderOrder | `COUNTRY_BOUNDARY_CONFIG.hoverLineRenderOrder` | `2.3` | Hover line level |
 | Border hover glow opacity | `COUNTRY_BOUNDARY_CONFIG.hoverGlowOpacity` | `0.38` | Glow line opacity |
 | Border hover glow line width | `COUNTRY_BOUNDARY_CONFIG.hoverGlowLineWidth` | `3` | Glow `LineBasicMaterial.linewidth` |
 | Border hover glow level offset | `COUNTRY_BOUNDARY_CONFIG.hoverGlowRenderOrderOffset` | `0.01` | Glow renderOrder = `2.29` |
-| Border hover glow radius offset | `COUNTRY_BOUNDARY_CONFIG.hoverGlowRadiusOffset` | `0.04` | Glow radius = hover radius + 0.04 |
+| Border hover glow radius offset | `COUNTRY_BOUNDARY_CONFIG.hoverGlowRadiusOffset` | `0` | Glow uses the same radius as the hover line to avoid coastline detail misalignment |

 ## Real Terrain

@@ -160,10 +168,13 @@ The land/ocean base is an Earth base-map asset and preloads at startup; the "Bor

 | Name | Variable | Current Value | Location / Notes |
 | --- | --- | --- | --- |
-| Satellite display radius offset | `SATELLITE_CONFIG.displayAltitudeOffset` | `8` | Satellite point position |
+| Satellite fallback radius offset | `SATELLITE_CONFIG.fallbackAltitudeOffset` | `8` | Legacy same-sphere position when real altitude is disabled or TLE propagation fails |
+| Satellite altitude compression scale | `SATELLITE_CONFIG.altitudeCompressionKm` | `1200` | TLE/SGP4 altitude compression; larger values soften high-orbit separation |
+| Satellite real-altitude cap | `SATELLITE_CONFIG.maxDisplayAltitudeKm` | `40000` | Altitude clamp for the compressed display mapping, covering GEO-range altitude |
+| Satellite real-altitude display offset range | `minRealAltitudeOffset / maxRealAltitudeOffset` | `4 / 25` | Compressed offset added to `CONFIG.earthRadius`; `25` is about one quarter of the current globe radius `100`, keeping high-orbit separation visible without pushing GEO/MEO trails too far away |
 | Satellite dot base pixel size | `SATELLITE_CONFIG.dotBaseSize` | `2.8` | Point shader size |
 | Satellite backdrop dot scale | `SATELLITE_CONFIG.dotBackdropScale` | `1.28` | Backdrop dot size |
-| Satellite dot opacity range | `dotOpacityMin / dotOpacityMax` | `0.7 / 1.0` | Breathing animation |
+| Satellite dot opacity range | `dotOpacityMin / dotOpacityMax` | `0.42 / 1.0` | Breathing animation |
 | Satellite dot breathing speed | `SATELLITE_CONFIG.dotBreathingSpeed` | `0.12` | Dot opacity animation |
 | Satellite backdrop renderOrder | inline | `5` | `satelliteBackdropPoints.renderOrder` |
 | Satellite dot renderOrder | inline | `6` | `satellitePoints.renderOrder` |
@@ -230,7 +241,9 @@ AIS vessel markers use batched `THREE.Points`, not one `THREE.Sprite` per vessel
 | Medium color | `BGP_CONFIG.severityColors.medium` | `0xffd166` | Medium-severity event |
 | Low color | `BGP_CONFIG.severityColors.low` | `0x4dabf7` | Low-severity event |
 | Collector base color | `BGP_CONFIG.collectorColor` | `0x6db7ff` | Default collector color |
-| Region color | `BGP_CONFIG.regionColor` | `0x2dd4bf` | Region overlay |
+| BGP halo neutral tint | `BGP_CONFIG.halo.tintNeutralColor` | `0xffffff` | Broad halos derive from the matching event / collector icon color and blend toward this neutral, keeping the same hue family without becoming identical |
+| BGP halo icon-color blend | `BGP_CONFIG.halo.tintBlend` | `0.72` | Higher values stay closer to the icon color; currently used by event region halos, collector coverage halos, and collector radar pulse |
+| Region color | `BGP_CONFIG.regionColor` | `0x2dd4bf` | Legacy region baseline; broad BGP halos now prefer colors derived from the event or collector icon |

 ## Celestial and Starfield

--- a/docs/technical/en/earth-news-live-streams-collector-format.md
+++ b/docs/technical/en/earth-news-live-streams-collector-format.md
@@ -103,7 +103,7 @@ Goals:

 ## Collector Configuration

-`news_live_streams` does not need a separate new page; it reuses Collector Settings under `/settings`:
+`news_live_streams` does not need a separate new page; it reuses Collectors under `/collection-management`:

 - `endpoint`
  - Channel directory JSON API URL
--- a/docs/technical/en/earth-render-layer-order.md
+++ b/docs/technical/en/earth-render-layer-order.md
@@ -6,8 +6,8 @@ Note: the layer control panel order and the registration / startup load order ar

 | Order type | Current sequence | Notes |
 | --- | --- | --- |
-| Control panel order | Cables → Trails → Satellites → Compute Centers → BGP → Terrain → HD Texture → Cloud Layer → Border Lines → Grid | Controlled by `displayOrder`, sorted by operational relevance. |
-| Registration / startup load order | Grid → Border Lines / Land-Ocean Base → HD Texture → Cloud Layer → Cables → Compute Centers → BGP → Satellites | Controlled by registration order and `startupPriority`, sorted surface-to-sky; the startup queue reads persisted layer visibility first, skips normal layers explicitly saved as hidden, and HD Texture does not download the texture when disabled; Border Lines are the exception: the land-ocean base always preloads, while the persisted state only controls interactive border lines and hover; Trails and Terrain are dependency/optional display layers and do not participate in normal startup data loading. |
+| Control panel order | Cables → Satellites → Compute Centers → Vessels → BGP → Terrain → HD Texture → Cloud Layer → Border Lines → Grid | Controlled by `displayOrder`, sorted by operational relevance; satellite trails moved to Settings and are no longer a layer-list item. |
+| Registration / startup load order | Grid → Border Lines / Land-Ocean Base → HD Texture → Cloud Layer → Cables → Compute Centers → Vessels → BGP → Satellites | Controlled by registration order and `startupPriority`, sorted surface-to-sky; the startup queue reads persisted layer visibility first, skips normal layers explicitly saved as hidden, and HD Texture does not download the texture when disabled; Border Lines are the exception: the land-ocean base always preloads, while the persisted state only controls interactive border lines and hover; Trails and Terrain are dependency/optional display layers and do not participate in normal startup data loading. |

 ## Surface Layer Stack

@@ -17,8 +17,8 @@ Note: the layer control panel order and the registration / startup load order ar
 | -1 | Earth occluder sphere | `earth.js` | Invisible inner sphere | Writes depth buffer | Occludes objects behind the Earth. |
 | 0 | Earth base sphere | `earth.js` | `CONFIG.earthRadius` | Surface picking fallback target | Dark base; still visible when all optional map layers are off. |
 | 0.2 | Country dark tint | `country-boundaries.js` | `tintAltitudeOffset` | Raycast disabled | Used when HD texture is off. |
-| 0.86 | Land/ocean base fill | `country-boundaries.js` | `landAltitudeOffset`; ocean `#010609`, land `#080f1b` | Raycast disabled | Base map remains usable even when country borders are off. |
-| 0.96 | HD Earth texture | `earth.js` | `textureOverlayAltitudeOffset` | Surface picking target when visible | HD texture always overlays the land/ocean base fill. |
+| 0.86 | Land/ocean base fill | `country-boundaries.js` | `landAltitudeOffset = 0.32`; ocean `#010609`, land `#080f1b` | Raycast disabled | Base map remains usable even when country borders are off; radius is separated from the base sphere to avoid far-zoom z-fighting. |
+| 0.96 | HD Earth texture | `earth.js` | `textureOverlayAltitudeOffset = 0.48` | Surface picking target when visible | HD texture always overlays the land/ocean base fill; radius must stay above the land/ocean base and far enough from the base sphere. |
 | 1 | Atmospheric glow and clouds | `earth.js` | Atmosphere / cloud spheres | Not in normal object selection path | Cloud layer controlled by the "Cloud Layer" toggle. |
 | 1 | Submarine cables | `cables.js` | `CABLE_CONFIG.line.renderOrder` | Cable picking path | Preserves existing cable layer level. |
 | 1.2 | Real terrain | `earth.js`, `terrain.js` | `TERRAIN_CONFIG.baseRadiusOffset` plus terrain displacement | Raycast disabled | Terrain overlays HD texture; temporarily hidden when HD texture is off, restores to prior state when re-enabled. |
@@ -29,9 +29,9 @@ Note: the layer control panel order and the registration / startup load order ar
 | 3 | Satellite footprint fill / Iridium coverage ring | `satellites.js`, `iridium-footprint-adapter.js` | `GROUND_FOOTPRINT_RENDER_ORDER` | depth-tested; Iridium adapter fill / ring use the same renderOrder | Footprint above land / texture / terrain and country borders, below compute centers and satellites. |
 | 3-5 | BGP markers and overlays | `bgp.js` | Each marker's own renderOrder | BGP picking path | Preserves existing BGP visual level. |
 | 4.5 | Compute centers | `compute-centers.js` | `COMPUTE_CENTER_RENDER_ORDER` | Compute center picking path | Surface facilities, below satellites. |
-| 5 | Satellite background dot | `satellites.js` | Fixed renderOrder | Screen-space satellite picking | Below satellite dots. |
-| 6 | Satellite dots | `satellites.js` | Fixed renderOrder | Screen-space satellite picking | Satellite dots above footprints and compute centers. |
-| 12+ | Satellite locked ring, halo, predicted orbit | `satellites.js` | `SATELLITE_CONFIG.overlayRenderOrder` and offsets | Satellite overlay path | Used for selected/locked satellite emphasis. |
+| 5 | Satellite background dot | `satellites.js` | Fixed renderOrder; by default TLE/SGP4 altitude is compressed to `CONFIG.earthRadius + 4..25`; with real altitude disabled or propagation failed, uses `fallbackAltitudeOffset = 8` | Screen-space satellite picking | Below satellite dots. |
+| 6 | Satellite dots | `satellites.js` | Same compressed / fallback height as satellite backdrop dots | Screen-space satellite picking | Satellite dots above footprints and compute centers. |
+| 12+ | Satellite locked ring, halo, predicted orbit | `satellites.js` | `SATELLITE_CONFIG.overlayRenderOrder` and offsets; predicted orbit follows the same real-altitude toggle and fixes the lock-time globe pose to draw a closed inertial orbit; returns to same-sphere mode when real altitude is disabled | Satellite overlay path | Used for selected/locked satellite emphasis. |
 | 98-100 | Sun / moon halo and sprite | `celestial.js` | Fixed renderOrder | Celestial picking disabled | Foreground celestial sprites. |

 ## Toggle Behavior
@@ -43,6 +43,18 @@ Note: the layer control panel order and the registration / startup load order ar
 | Terrain on | Displayed above HD texture, but below country border hover, footprints, satellites, and other emphasis layers. |
 | Cloud layer | Only controls cloud mesh visibility. |
 | Border Lines off | Hides only interactive border lines and hover, clearing hover state; the land/ocean base fill remains as the Earth base map. |
+| Real Satellite Altitude off | Satellite dots, trails, and predicted orbits use the legacy same-sphere height; satellites with missing TLE data or failed propagation also use this fallback height. |
+
+## Depth Spacing Rules
+
+The Earth surface is not a single mesh. It is a stack of near-concentric shells: base sphere, land/ocean base, HD texture, terrain, clouds, atmosphere, and the occluder. Radius offsets that look harmless at close zoom can collapse into the same depth-buffer pixels at zoomed-out views such as 50%, causing z-fighting that appears as black blocks, snow, or flicker.
+
+Maintenance rules:
+
+- Do not reach first for hiding layers at far zoom. Check neighboring shell `altitudeOffset`, `renderOrder`, `depthTest`, and `depthWrite` first.
+- Whole-globe overlays such as the land/ocean base and HD texture must keep explicit separation from `CONFIG.earthRadius`; the current stable values are `landAltitudeOffset = 0.32` and `textureOverlayAltitudeOffset = 0.48`.
+- Any new whole-globe or near-whole-globe surface overlay must be screenshot-verified at 50% zoom and at common close zooms, with no black blocks, snow, flicker, or obvious floating.
+- If these radii change, update this document and the intent around the constants in `frontend/public/earth/js/constants.js`.

 ## Interaction Rules

--- a/docs/technical/en/faq.md
+++ b/docs/technical/en/faq.md
@@ -55,12 +55,12 @@ tasklist /svc /fi "PID eq 4700"
 For temporary troubleshooting, you can stop IP Helper from Administrator PowerShell:

 ```powershell
-Stop-Service iphlpsvc
+Stop-Service iphlpsvc -Force
 ```

-This may affect networking, proxying, or forwarding features. Do not disable it long-term unless you know why it is safe. If the Windows forwarding rule must stay, use a different Planet backend port.
+This may affect networking, proxying, or forwarding features. Do not disable it long-term unless you know why it is safe. `--allow-lan` now binds `3000` / `8000` / `8010` directly, so persistent portproxy is no longer required.

-If the script prints `failed-stop-service` or `failed-stop-process`, the current shell does not have permission to clear the Windows listener. Startup stops immediately instead of launching the backend into the same port conflict.
+If the script prints `Windows 侧端口 ... 存在监听者`, or Vite reports `Port 3000 is already in use` followed by `Windows listener ... services=iphlpsvc`, an old Windows listener still owns the port. The script requests Administrator PowerShell cleanup for that port. If the automatic cleanup is canceled, inspect `netsh interface portproxy show all`, delete the matching `listenport` rule, confirm the PID and services with `netstat` / `tasklist` if no portproxy rule exists, and temporarily run `Stop-Service iphlpsvc -Force` when appropriate. After old rules are gone, rerun `./planet.sh restart --allow-lan`; LAN devices still use `3000` / `8000` / `8010`.

 ### Which startup flags change default ports?

@@ -87,6 +87,7 @@ Check in this order before changing firewall rules:
 # In WSL or the shell running Planet
 curl http://localhost:3000
 curl http://localhost:8000/health
+curl http://localhost:8010/health
 ```

 Then verify from Windows PowerShell:
@@ -94,6 +95,7 @@ Then verify from Windows PowerShell:
 ```powershell
 curl http://localhost:3000
 curl http://localhost:8000/health
+curl http://localhost:8010/health
 ```

 If both localhost checks pass but a phone or another computer cannot connect, start with LAN enabled:
@@ -108,22 +110,28 @@ The flag must be written as `--allow-lan`. `allowlan` or `--allowlan` is not rec
 ./planet.sh restart -f 3000 --allow-lan
 ```

-If `ss -ltnp` shows the frontend listening on `0.0.0.0:3000`, but `Test-NetConnection <Windows LAN IP> -Port 3000` still fails from Windows PowerShell, the problem is usually Windows-side forwarding or firewall policy rather than Vite or `.zshrc`.
+If `ss -ltnp` shows the frontend listening on `0.0.0.0:3000`, but `Test-NetConnection <Windows LAN IP> -Port 3000` still fails from Windows PowerShell, the problem is usually Windows-side port ownership, stale `portproxy`, or firewall policy rather than Vite or `.zshrc`.

-For traditional WSL NAT networking, configure portproxy and firewall from Administrator PowerShell:
+`./planet.sh start --allow-lan` directly exposes `3000` / `8000` / `8010` and checks port availability, stale `portproxy`, and Windows Firewall before startup. If a Windows-side listener owns a port, the script requests Administrator PowerShell cleanup. When inbound allow rules are missing, it also triggers a UAC Administrator PowerShell request to create them. If the automatic request is canceled, clean up manually:

 ```powershell
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=3000 connectaddress=127.0.0.1 connectport=3000
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=8000 connectaddress=127.0.0.1 connectport=8000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=3000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8010
+
 New-NetFirewallRule -DisplayName "WSL Planet 3000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3000
 New-NetFirewallRule -DisplayName "WSL Planet 8000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8000
+New-NetFirewallRule -DisplayName "WSL Planet 8010" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8010
 ```

+LAN devices should use the Windows external port, for example `http://<Windows LAN IP>:3000/earth`.
+
 If `wslinfo --networking-mode` prints `mirrored`, also check Hyper-V firewall. Even when ordinary Windows Firewall rules exist, Hyper-V firewall can still block external devices from reaching WSL. From Administrator PowerShell, allow the required ports:

 ```powershell
 New-NetFirewallHyperVRule -Name "Planet-Frontend-3000" -DisplayName "Planet Frontend 3000" -Direction Inbound -VMCreatorId "{40E0AC32-46A5-438A-A0B2-2B479E8F2E90}" -Protocol TCP -LocalPorts 3000 -Action Allow
 New-NetFirewallHyperVRule -Name "Planet-Backend-8000" -DisplayName "Planet Backend 8000" -Direction Inbound -VMCreatorId "{40E0AC32-46A5-438A-A0B2-2B479E8F2E90}" -Protocol TCP -LocalPorts 8000 -Action Allow
+New-NetFirewallHyperVRule -Name "Planet-AIProvider-8010" -DisplayName "Planet AI Provider 8010" -Direction Inbound -VMCreatorId "{40E0AC32-46A5-438A-A0B2-2B479E8F2E90}" -Protocol TCP -LocalPorts 8010 -Action Allow
 ```

 Use these commands to inspect the current Hyper-V firewall state:
@@ -286,7 +294,7 @@ If only AI Provider is unhealthy, restart just that service:

 ### Connectivity validation passes, but collection cannot read credentials. Why?

-Connectivity validation can read saved console settings, environment variables, and some credentials from `~/.zshrc`. For actual collection, prefer saving credentials in Settings -> Collector Settings, especially for AISStream's long-lived WebSocket collector.
+Connectivity validation can read saved console settings, environment variables, and some credentials from `~/.zshrc`. For actual collection, prefer saving credentials in Collection Management -> Collectors, especially for AISStream's long-lived WebSocket collector.

 If `AISSTREAM_API_KEY` only lives in `~/.zshrc`, confirm the backend process actually inherited it. Otherwise validation may pass while the collector runtime has no key.

@@ -300,7 +308,7 @@ export BARENTSWATCH_CLIENT_ID="..."
 export BARENTSWATCH_CLIENT_SECRET="..."
 ```

-For stable operation, save credentials in Collector Settings so connectivity validation, collection jobs, and Earth realtime aggregation use the same configuration.
+For stable operation, save credentials in Collectors so connectivity validation, collection jobs, and Earth realtime aggregation use the same configuration.

 ## Docs / Permissions

@@ -315,12 +323,37 @@ Docs visibility is controlled by Gatekeeper groups:

 ## Earth Common Tasks

+### Why does Earth say the boundary endpoint is not configured, or only show low precision boundaries?
+
+Country boundaries have moved out of the collector system. They are no longer generated by datasource collection tasks. The low-precision boundary file is bundled with the frontend and is the expected fallback when no local high-precision PMTiles artifact exists.
+
+There are two high-precision entry points:
+
+- Earth page settings gear -> Boundary Precision: switching to High Precision starts the first background download/build, shows percentage progress, and applies the result automatically.
+- Console `Operations and Configuration -> Earth Content -> Boundary Precision`: use this to inspect provider, manifest, PMTiles, fallback state, edit source JSON, or rebuild manually.
+
+If the UI says the update source is incomplete, save the source configuration from `Earth Content -> Boundary Precision`. The private local config is written to `config/earth-boundary-sources.local.json`; do not commit it. Falling back to low precision is normal when no high-precision artifact has been built.
+
 ### Why did collecting a location candidate not write anything?

 Collecting and saving are two separate actions. Candidates can be previewed on Earth first. A candidate is written only after clicking Save or using the unresolved list's one-click adopt flow.

 Compute-center saves write to `compute_center_locations` and refresh the layer. Records with no candidate stay in the unresolved list; Planet does not fabricate a location from a country center or hard-coded hint.

+### Why did the Earth logo or title not return to the default after I edited it?
+
+Earth brand assets are managed from `Operations and Configuration -> Earth Content -> Brand Assets`. Uploaded images are stored as Earth brand asset URLs. If text fields such as title or ARIA label are cleared, Planet falls back to default text so the HUD never renders an empty brand.
+
+Use `Reset Brand Assets` to restore the shipped logo, title image, and copy. Refreshing Earth does not delete the saved runtime brand configuration.
+
+### Why are satellites no longer on one sphere?
+
+Earth enables "Real Satellite Altitude" by default. Satellite positions still come from TLE/SGP4, but altitude is compressed for display: LEO satellites stay close to the globe, while higher-orbit satellites render farther out without leaving the normal view. The maximum display offset is `25`, about one quarter of the current globe radius; this is a readability compromise that separates GEO / MEO / LEO without drawing real kilometers to scale. This setting also affects satellite trails and the predicted orbit shown after locking a satellite.
+
+Turn off "Real Satellite Altitude" in Earth Settings to restore the legacy same-sphere satellite display. Satellites with missing TLE data or failed propagation still fall back to the legacy fixed height, so they do not disappear just because a real altitude cannot be computed.
+
+Low-inclination high-orbit satellites should stay near the equator or a fixed longitude band, not form a near north-south ring. Earth converts the SGP4 inertial position to Earth-fixed coordinates for the current dot; the locked predicted orbit fixes the current globe pose and draws one inertial orbit, so it should close and preserve the correct inclination. If a future details card shows inclination near `0°` but the predicted orbit looks polar, check the ECI/TEME-to-ECF conversion and whether the predicted orbit accidentally uses per-sample `gstime`.
+
 ### Why does Motion Debug not show camera video?

 With the Browser Camera source, the debug panel shows the local browser camera preview and draws the skeleton over it. If `Skeleton Only` is enabled, the video preview is hidden and the panel keeps only the dark canvas plus red/green skeleton.
--- a/docs/technical/en/frontend-admin-frontend-context.md
+++ b/docs/technical/en/frontend-admin-frontend-context.md
@@ -33,6 +33,8 @@ Current admin-related routes:
 - `/alerts/situational`
 - `/bgp`
 - `/ai`
+- `/earth-content`
+- `/collection-management`
 - `/settings`

 `/earth` is a standalone display page and is not part of the console shell.
@@ -277,6 +279,42 @@ Constraints:
 - Do not let tables blow out the full page
 - New table areas should reuse `TableScrollRegion` / `ScrollbarOverlay`

+### Datasource Directory Page
+
+[DataSources.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataSources/DataSources.tsx) is the datasource directory and collection operation page. It should not grow back into a configuration editor.
+
+Current page boundary:
+
+- Built-in and custom sources are merged as `UnifiedDataSource`.
+- The list shows type, state, last run, collection progress, and actions.
+- Clicking a name opens a read-only drawer.
+- Endpoint, headers, and config are displayed here, not edited.
+- Credential-bearing collectors point users to `Collection Management -> Collectors`.
+
+Keep this boundary: do not put custom datasource editing, built-in endpoint overrides, or credential forms back into `/datasources`. Those configuration entry points live at `/collection-management?tab=collector_credentials`.
+
+### Collectors Page
+
+[Settings.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Settings/Settings.tsx) has three route modes: `/settings` for System Settings, `/earth-content` for Earth Content, and `/collection-management` for Collection Management. The `collector_credentials` tab is shown as `Collectors` under `/collection-management`.
+
+Current boundary:
+
+- The dropdown selects built-in collectors.
+- The plug icon beside the dropdown runs the health check.
+- Credential-bearing collectors place credential forms above base config.
+- Free collectors show endpoint, default endpoint, headers, timeout, and retry.
+- BarentsWatch AIS keeps its dedicated credential form.
+
+### Earth Content Page
+
+`/earth-content` reuses the same single-screen tab container from [Settings.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Settings/Settings.tsx), but its ownership is separate from System Settings:
+
+- `TV Livestream` owns the Earth media-panel source configuration.
+- `Boundary Precision` owns the Earth static boundary asset state: provider, low-precision fallback, high-precision manifest/PMTiles, source JSON, and build action.
+- `Base Map`, `Layer Resources`, `3D Assets`, and `News Anchor Strategy` are placeholders only. They show module status and do not invent fake APIs or fake data.
+
+Do not add Earth experience resources or collection-lifecycle tabs back into `/settings`; collection belongs to `/collection-management`, and Earth display resources belong to `/earth-content`.
+
 ### 3. Complex Workspace Pages

 Examples:
--- a/docs/technical/en/manual.md
+++ b/docs/technical/en/manual.md
@@ -76,15 +76,17 @@ The console at `http://localhost:3000/admin` is built with React + Ant Design. T
 | BGP Alerts | `/alerts/bgp` | BGP-related alerts |
 | Situational Alerts | `/alerts/situational` | Situational analysis alerts |
 | AI | `/ai` | Model providers, tools, testbench |
+| Earth Content | `/earth-content` | TV livestreams, boundary precision, base-map and layer-resource entry points |
+| Collection Management | `/collection-management` | Collectors, scheduling, collection history entry points |
 | Logs | `/logs` | Usually visible only to super admin |
 | Users | `/users` | Create/delete users, change roles/groups |
-| Settings | `/settings` | System, SMTP, TV, collectors |
+| System Settings | `/settings` | Display, notification, security, SMTP |

 Menu items hide automatically when you lack permission. If a menu is missing, check your role and Gatekeeper groups.

 ## Configure Data Collectors

-`/settings?tab=collector_credentials` is the "Collector Settings" page. It manages connection configuration for every collector, not just credentials.
+`/collection-management?tab=collector_credentials` is the "Collectors" page. It manages connection configuration for every collector, not just credentials. Legacy `/settings?tab=collector_credentials` redirects here; the datasource directory remains at `/datasources`.

 Steps:

@@ -125,23 +127,25 @@ The default guide follows the BarentsWatch official tutorial and reminds you to

 Steps:

-1. Open `/settings?tab=collector_credentials` and select `AISStream Realtime Vessels : aisstream_vessels`
+1. Open `/collection-management?tab=collector_credentials` and select `AISStream Realtime Vessels : aisstream_vessels`
 2. Fill the AISStream API Key
 3. Keep the default endpoint `wss://stream.aisstream.io/v0/stream`
 4. Click the plug icon to test; confirm it reports `Reachable`
 5. Save collector settings
-6. Trigger the `aisstream_vessels` collector from the collection scheduler
-7. Watch the `AISStream Runtime` panel:
+6. Open the `Realtime Streams` tab on `/datasources` and find `AISStream Realtime Vessels`
+7. Use `Start`, `Stop`, or `Reconnect` there. The normal `Collection Tasks` tab does not count AISStream in one-click collection or percentage progress
+8. Watch the realtime stream panel:
   - `streaming` / `connected` means the live stream is being consumed
-   - `messages this round` should keep growing
-   - `disconnected` with `ConnectionResetError` means the upstream or network dropped; re-trigger or wait for reconnect
+   - `total stored`, `last 24h`, `last 1h`, and `unique MMSI` show historical collection volume
+   - `disconnected` with a recent error means the upstream or network dropped; click `Reconnect`

 ## Configure AI Credentials

-`/ai?tab=providers` is the AI management entry. Two key sub-tabs:
+`/ai?tab=providers` is the AI management entry. Three key sub-tabs:

 - `Model Providers`: default LLM provider, model, base URL, API key, local `aiprovider` proxy, connection test
 - `Tools`: a dropdown for specific tools — currently WebSearch and OCR
+- `Prompts`: a task dropdown for news localization, alert analysis, BGP briefs, and other LLM tasks. Operators can edit the prompt or reset it to the default

 ### Model Providers

@@ -162,6 +166,10 @@ The plug icon at the end of the Base URL input runs a connection test. A passing
 - **WebSearch**: provider, API key, base URL, max results, timeout, advanced provider parameters. While disabled, all fields except the enable switch are greyed out
 - **OCR**: provider, base URL, API key, model/engine, recognition languages, timeout, max file size, output format

+### Prompts
+
+After selecting a task, the page shows the effective prompt, whether it is customized, the shipped default version, and a reset button. Saving affects only that task. Reset restores the default prompt from the current release package. Business facts, context, and output schemas are still assembled by the backend for each task.
+
 The legacy link `/settings?tab=ai` redirects to `/ai?tab=providers`.

 ## System Settings
@@ -172,8 +180,27 @@ The legacy link `/settings?tab=ai` redirects to `/ai?tab=providers`.
 - **Notifications**: alert email switch, recipient, critical/warning/daily summary
 - **Security**: session timeout, max login attempts, password policy
 - **SMTP Email**: outgoing email used by registration and password reset (visible to `admin` / `super_admin` only)
- **TV Livestream**: TV source management
- **AI / WebSearch / OCR**: see above
+
+TV livestreams and boundary precision moved to `/earth-content`; collectors and scheduling moved to `/collection-management`; AI Provider / WebSearch / OCR live at `/ai`.
+
+### Earth Content
+
+`/earth-content` is under the console's Operations and Configuration group and owns resources used by the Earth frontend:
+
+- **Brand Assets**: manages the logo, title image, title text, subtitle, and description used by the Earth HUD. Uploaded images are saved as Earth brand assets and read by the Earth page immediately.
+- **TV Livestream**: manages sources shown in the Earth media panel.
+- **Boundary Precision**: shows the current provider, low-precision fallback, high-precision PMTiles/manifest status, local source JSON, and manual build action.
+- **Base Map**, **Layer Resources**, **3D Assets**, and **News Anchor Strategy**: placeholder tabs for future configuration. They do not display fake data.
+
+The Earth page settings gear also includes Boundary Precision. Switching to High Precision starts a local background download/build, like a game update package, when no high-precision asset exists yet. Progress is shown as a percentage, and the result applies automatically after success without a page reload. Switching back to Low Precision only changes the local display preference.
+
+### Collection Management
+
+`/collection-management` is also under Operations and Configuration and owns the collection lifecycle:
+
+- **Collectors**: endpoint, headers, credentials, timeout, retry, and connection checks.
+- **Collection Scheduling**: the existing scheduling configuration.
+- **Collection History / Snapshots**: a placeholder for future collection task, snapshot, and collected-data browsing.

 ### SMTP Email Settings

@@ -203,7 +230,7 @@ To let a regular user read developer or operations docs, add `docs_developer` or

 ## Data Exploration

- `/datasources`: source directory. It can be filtered by product domain, layer/module, enabled state, last run status, whether collected records exist, and search text. Selecting rows triggers only those sources; with no selected rows, `Collect current filter` triggers the filtered scope. Clicking a name opens an info drawer showing endpoint, headers, base config, and built-in flag; endpoint/credentials editing happens at `/settings -> Collector Settings`. The `Collecting N` tag under the overall progress can be clicked to expand the current collection task list
+- `/datasources`: source directory. The `Collection Tasks` tab is for one-shot, scheduled, and finite collectors; it can be filtered by product domain, layer/module, enabled state, last run status, whether collected records exist, and search text. Selecting rows triggers only those sources; with no selected rows, `Collect current filter` triggers the filtered scope. The `Realtime Streams` tab is for AISStream / WebSocket long connections and shows connection health, stored totals, time-window counters, and Start / Stop / Reconnect actions. Clicking a name opens an info drawer showing endpoint, headers, base config, and built-in flag; endpoint/credentials editing happens at `/collection-management -> Collectors`. The `Collecting N` tag under the overall progress can be clicked to expand the current collection task list
 - `/data`: collected data table — used to verify "did data arrive", "is the freshness right", "does a source emit valid records"
 - `/bgp`: BGP detail page with list + detail + analysis; complements the BGP layer on Earth
 - `/alerts/system`, `/alerts/bgp`, `/alerts/situational`: system, BGP, and situational alerts
@@ -228,7 +255,7 @@ A single globe view of: BGP events and observations, satellites and tracks, cabl

 ### Layer Control

-The right-side layer panel toggles layers. Common layers: graticule, country boundaries, high-res tiles, atmospheric clouds, cables, compute centers, BGP, satellites, AIS vessels, tracks, terrain.
+The right-side layer panel toggles layers. Common layers: graticule, country boundaries, high-res tiles, atmospheric clouds, cables, compute centers, BGP, satellites, AIS vessels, terrain. Satellite tracks are no longer listed as a standalone layer; they are controlled from Settings.

 Dependencies:

@@ -254,10 +281,16 @@ Candidates preview on Earth directly. Saving a compute-center candidate writes t

 ### Settings

-The settings panel covers: rotate / cruise / motion mode, cruise modules (BGP/news/compute centers/vessels/cables/satellites), view (satellite display style, day-night mode, panel toggles), motion debug mode / input source / skeleton-only, default globe size, terrain opacity, reset.
+The settings panel is grouped into Runtime, Display, Panels, Motion, Shortcuts, and System. It covers rotate / cruise / motion mode, cruise modules (BGP/news/compute centers/vessels/cables/satellites), view (satellite display style, hover tooltip, satellite idle breathing, real satellite altitude, track display, compact dots, day-night mode, panel toggles), motion debug mode / input source / skeleton-only, shortcut enablement and remapping, default globe size, terrain opacity, reset.
+
+"Real Satellite Altitude" is enabled by default: satellite positions use a compressed display height based on TLE/SGP4 orbital altitude. LEO satellites remain close to the globe, while high-orbit satellites render farther out without leaving the normal view. The high-orbit display height is capped at about one quarter of the globe radius, so GEO / MEO objects remain visually separated from LEO without spreading trails and selection targets too far apart. Turning it off restores the legacy same-sphere satellite display. "Track Display" controls satellite trail visibility; trails are unavailable while the satellite layer is hidden.
+
+"Hover Tooltip" controls the tooltip shown when the pointer hovers over the globe surface: `Country` shows country details only when land matches a country, and stays silent over oceans such as the Pacific; `Position` shows latitude, longitude, and elevation over land and ocean; `Full` is the default and shows country + position over land and position over ocean.

 These settings live in browser local storage; switching browsers or clearing site data resets them.

+Shortcuts also live in browser local storage. Use the Shortcuts category to disable an individual shortcut, capture a new key, or restore the default binding for the current browser only.
+
 ### View Controls

 | Action | Effect |
@@ -309,6 +342,7 @@ Mobile uses a drawer layout: layer control moves into a drawer; search/settings/
 - **Earth does not open**: confirm the frontend is online; if not on port `3000`, use the port printed by the startup log
 - **Layers have no data**: open `/datasources` to check source status, collected-record state, and the latest run result; then `/data` or `/bgp` for records
 - **Satellites / BGP / cables load slowly**: those layers depend on backend APIs and external data sources; the first load waits for startup tasks
+- **Satellites are not all on one shell**: this is the default compressed real-altitude display. Disable "Real Satellite Altitude" in Settings to return to the legacy same-sphere view

 ## Docs Site

--- a/docs/technical/en/ops-planet-sh-startup.md
+++ b/docs/technical/en/ops-planet-sh-startup.md
@@ -22,10 +22,10 @@ image_exists AND stamp_non_empty AND fingerprint_match

 ### Fix

-The stamp file moved to a persistent cache path:
+The stamp file moved from a temporary location to a persistent cache path:

 ```bash
-AI_PROVIDER_BUILD_STAMP_FILE="$HOME/.cache/planet/aiprovider_build.sha256"
+AI_PROVIDER_BUILD_STAMP_FILE="${XDG_CACHE_HOME:-$HOME/.cache}/planet/aiprovider_build.sha256"
 ```

 Writing the stamp creates the directory first:
@@ -92,7 +92,7 @@ COPY aiprovider /app/aiprovider

 ### Runtime Configuration

-Before starting AI Provider, `planet.sh` generates a temporary env-file and passes it to Compose or the manual `docker run` fallback. Configuration priority:
+Before starting AI Provider, `planet.sh` generates a current-user runtime env-file and passes it to Compose or the manual `docker run` fallback. The default path is `${XDG_STATE_HOME:-$HOME/.local/state}/planet/aiprovider_runtime.env`. Configuration priority:

 1. `aiprovider/.env`
 2. simple `export AI_...=...` or `AI_...=...` lines from `~/.zshrc`
@@ -177,7 +177,7 @@ Frontend startup now has an additional pre-start cleanup retry layer:
 - `PORT_PRESTART_RETRIES`: defaults to 3 attempts.
 - `PORT_PRESTART_RETRY_INTERVAL`: defaults to 2 seconds.

-`kill_port_if_requested()` first cleans listener PIDs visible in the current environment. It only checks for Windows-side listeners when the script detects WSL, no local listener PID is visible, and the port still cannot bind. In that WSL-only path it tries to stop the owning Windows service or force-stop the owning process through PowerShell. If permissions are missing, or a system service such as `iphlpsvc` refuses to stop, the script prints the Windows listener details and stops startup immediately instead of launching the service into the same port error. Non-WSL environments do not run the Windows cleanup path. At that point, use Administrator PowerShell to clear the portproxy/service ownership, or choose another port.
+`kill_port_if_requested()` first cleans listener PIDs visible in the current environment. It only checks for Windows-side listeners when the script detects WSL, no local listener PID is visible, and the port still cannot bind. In that WSL-only path it requests Administrator PowerShell to delete stale `portproxy` rules, stop services that own the port, or force-stop the owning process. If the administrator request is canceled, or a system service such as `iphlpsvc` refuses to stop, the script prints the Windows listener details and Administrator PowerShell recovery commands, then stops startup immediately instead of launching the service into the same port error. If the frontend Vite process only discovers `Port 3000 is already in use` after launch, the script prints the same Windows listener recovery commands. Non-WSL environments do not run the Windows cleanup path. `--allow-lan` now exposes `3000` / `8000` / `8010` directly and no longer starts an extra Windows forwarding process; old persistent portproxy rules should be removed.

 ## Issue 4: `restart` Behavior

@@ -188,6 +188,62 @@ Before the stamp path fix:

 After moving the stamp file, plain `restart` uses the same `stop + start` behavior and the same fingerprint check as `restart -b`.

+## State Files, Logs, and Failed-Start Cleanup
+
+`planet.sh` no longer writes PID files, logs, or runtime env-files to fixed `/tmp/planet_*` paths. The default state directory is:
+
+```bash
+${XDG_STATE_HOME:-$HOME/.local/state}/planet
+```
+
+At startup the script creates this directory and tries to set it to `700`. The current files include:
+
+- `backend.pid` / `frontend.pid` / `motion_agent.pid`
+- `backend.log` / `frontend.log` / `motion_agent.log`
+- `aiprovider_build.log`
+- `aiprovider_runtime.env`
+- `ports.env`
+
+PID writes validate that the PID is a positive integer, include a trailing newline, and try to set file mode `600`. PID reads ignore invalid content instead of passing it to `kill`.
+
+After a successful `start`, the script records the ports in `ports.env`. Later `./planet.sh health` calls prefer the last started ports; if the state file is missing, health checks fall back to the defaults `8000`, `3000`, `8010`, and `8765`. This avoids checking default ports after starting with custom ports.
+
+Startup now has light failed-start cleanup. If `start` exits before completing, the script only cleans local processes that this run already started: backend, frontend, and Motion Agent. It does not stop services after a successful start. AI Provider, PostgreSQL, and Redis keep their existing container lifecycle behavior.
+
+## Health Checks and Hardening
+
+HTTP readiness checks now use `curl -fsS --max-time`, so 4xx and 5xx responses are no longer treated as healthy.
+
+Process termination now validates:
+
+- signal names are limited to `TERM`, `KILL`, `INT`, and `HUP`;
+- PIDs must be positive integers;
+- process group IDs must be positive integers.
+
+This prevents bad PID files or invalid signals from reaching `kill`.
+
+Frontend and Motion Agent startup failures now call `print_port_listener_details()`, matching backend port diagnostics. The Windows-side listener and cleanup path still only runs when WSL is detected.
+
+## Cross-Platform Notes
+
+The script is currently Linux-first with WSL enhancements. Normal Linux runs do not execute the PowerShell path; WSL gets extra Windows listener, portproxy, and camera guidance.
+
+To make this single script fully portable across Linux, macOS, and WSL, the remaining platform differences should be wrapped behind compatibility helpers:
+
+- `stat --format`, `sort -V`, and `xargs -r` are GNU-style and are not fully compatible with default macOS BSD tools.
+- `hostname -I`, `ss`, `fuser`, and `systemctl` are usually unavailable on macOS.
+- `tac` may be missing on macOS; use `awk` or Python as a fallback.
+- Docker Desktop on macOS does not use `systemctl` daemon diagnostics.
+- Camera auto-detection relies on `/dev/video*` / `v4l2-ctl`, which is Linux-specific; macOS should use explicit camera URLs or a separate AVFoundation detector.
+
+The recommended direction is a small platform compatibility layer for port listener detection, version comparison, file metadata, reverse tail, LAN IP discovery, and Docker daemon diagnostics, instead of scattering more platform branches throughout service startup logic.
+
+## Production Delivery Boundary
+
+`planet.sh` is a local development convenience script, not the production startup entrypoint. Production delivery should use Kubernetes `Deployment`, `Service`, `Ingress`, and readiness/liveness probes for ports, health checks, restarts, and rolling upgrades. This removes the need for a host script to reclaim local ports and avoids running the Vite dev server in production.
+
+The production frontend shape is `vite build` static output served by nginx/Caddy or an equivalent HTTP server. Do not use `bun run dev` or `vite preview` in production. The project does not maintain a parallel Webpack build chain; if a future enterprise requirement needs closer Webpack-ecosystem compatibility, run an Rsbuild/Rspack spike first. Electron should only be evaluated when the official target becomes an offline desktop application.
+
 ## Optional Motion Agent Startup

 `planet.sh` can now manage the local Motion Capture Agent. It is disabled by default so ordinary development machines do not fail startup when cameras, OpenCV, or MediaPipe are unavailable.
--- a/docs/technical/en/ops-runbook.md
+++ b/docs/technical/en/ops-runbook.md
@@ -120,7 +120,7 @@ Useful for:
 - Demoing Earth from a phone or tablet
 - Other LAN machines reaching the same dev instance

-`--allow-lan` only makes the frontend and backend listen on `0.0.0.0`. When Planet runs in WSL, Windows can usually reach it through `localhost`, but other LAN machines hitting `http://<Windows LAN IP>:3000` still need Windows port forwarding and firewall rules.
+`--allow-lan` directly exposes the frontend, backend, and AI Provider from the development machine: frontend `3000`, backend `8000`, and AI Provider `8010`. Before startup, the script checks all three ports. If WSL/Linux cannot release a port and a Windows-side listener or stale `portproxy` rule owns it, the script requests Administrator PowerShell cleanup. When Planet runs in WSL, Windows can usually reach it through `localhost`; other LAN machines reaching the Windows LAN IP still need Windows Firewall allow rules.

 Diagnose in this order:

@@ -128,19 +128,24 @@ Diagnose in this order:
 # From the shell running Planet
 curl http://localhost:3000
 curl http://localhost:8000/health
-ss -ltnp | grep -E ':3000|:8000'
+curl http://localhost:8010/health
+ss -ltnp | grep -E ':3000|:8000|:8010'
 ```

-If WSL shows `0.0.0.0:3000` / `0.0.0.0:8000` but the LAN IP still fails, configure Windows from an elevated PowerShell:
+If the services are running but the LAN IP still fails, first remove stale `portproxy` rules and confirm Windows Firewall allows the ports. The script checks this automatically and requests Administrator PowerShell when needed. Manual fallback commands:

 ```powershell
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=3000 connectaddress=127.0.0.1 connectport=3000
-netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=8000 connectaddress=127.0.0.1 connectport=8000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=3000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8010

 New-NetFirewallRule -DisplayName "WSL Planet 3000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3000
 New-NetFirewallRule -DisplayName "WSL Planet 8000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8000
+New-NetFirewallRule -DisplayName "WSL Planet 8010" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8010
 ```

+LAN devices should use the Windows external ports, for example `http://<Windows LAN IP>:3000/earth`, `http://<Windows LAN IP>:8000/health`, and `http://<Windows LAN IP>:8010/health`.
+
 ## AI Provider Environment and Builds

 AI Provider runtime configuration lives in two places:
@@ -236,6 +241,15 @@ uv sync
 uv run pytest backend/tests/test_otp_service.py
 ```

+## Earth Boundary PMTiles Operations
+
+1. In the console, open `Operations and Configuration -> Earth Content -> Boundary Precision` to save boundary source configuration. The local config is written to `config/earth-boundary-sources.local.json`; do not commit it.
+2. Click "Build high precision boundaries", or switch the Earth toolbar settings gear to High Precision for the first build. The backend downloads the three source packages to `data/earth-boundary-sources/`, writes the source manifest, and invokes the PMTiles build script.
+3. The builder requires `tippecanoe` and `pmtiles` on PATH. Missing tools return a clear API error and do not write data-source collection records.
+4. A successful production build outputs `frontend/public/earth/data/boundaries/earth-boundaries-china-pov-v1.pmtiles` and its manifest.
+5. After deployment, open Earth, enable "Border Lines", and inspect China's southeast coast, Taiwan, Hainan, the South China Sea, Zangnan, Kosovo, and Gaza for hover behavior and boundary policy.
+6. If no high-precision manifest/PMTiles exists locally, Earth uses the bundled `frontend/public/earth/data/countries-admin0.min.geojson` fallback. If high-precision assets exist but tile requests fail, troubleshoot PMTiles range requests, manifest provider, Nginx `.pmtiles` static serving, and sha256 consistency.
+
 ## Related Docs

 - [planet.sh Startup Mechanism](/home/ray/dev/linkong/planet/docs/technical/en/ops-planet-sh-startup.md)
--- a/docs/technical/en/quickstart.md
+++ b/docs/technical/en/quickstart.md
@@ -11,7 +11,7 @@ Open the URL your administrator gave you, e.g. `http://planet.example.com`. A lo
 Entry points are split in two:

 - Public: `/earth` (3D situational view), `/docs` (public documentation)
- Login required: `/admin` (console), `/ai` (AI), `/settings` (system configuration)
+- Login required: `/admin` (console), `/ai` (AI), `/settings` (system settings), `/earth-content` (Earth content), `/collection-management` (collection management)

 ## 2. Register

@@ -32,9 +32,9 @@ The default role is `viewer`: you can sign in but only see public pages. For col

 After landing on the `/admin` dashboard, here's a recommended walk-through:

-1. `/settings?tab=collector_credentials`: pick a collector and click the plug icon to test connectivity. Free collectors (e.g. open BGP) usually work right away; credential-bearing ones like `AISStream` or `BarentsWatch` need an API key / client secret first
+1. `/collection-management?tab=collector_credentials`: pick a collector and click the plug icon to test connectivity. Free collectors (e.g. open BGP) usually work right away; credential-bearing ones like `AISStream` or `BarentsWatch` need an API key / client secret first
 2. `/ai?tab=providers`: fill an LLM provider (e.g. `minimax` / `openai`), model, base URL, API key, and click the plug at the end of the base URL to test. WebSearch / OCR tools are optional
-3. `/datasources` or `/data`: check whether the collectors have produced data
+3. `/datasources` or `/data`: check whether collectors have produced data. Use `/datasources -> Collection Tasks` for finite collectors, and `/datasources -> Realtime Streams` for AISStream / WebSocket health and counters
 4. `/alerts/system`: verify system alerts look right
 5. `/users` (super_admin only): open accounts for teammates or adjust their groups

@@ -48,7 +48,7 @@ Once in, verify:
 - Search finds cables, satellites, compute centers, BGP events
 - Compute-center and BGP collector detail cards can collect coordinate candidates and preview them on Earth
 - Mouse drag, wheel zoom, and the zoom percentage indicator work
- The settings panel can switch rotate / cruise / motion modes
+- The settings panel can switch rotate / cruise / motion modes; view settings can switch hover tooltip content, and satellite settings can toggle real-altitude layering and track display

 ## 5. Recover a Lost Password

--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
rayd1o	81970a1d05	release: bump version to 0.60.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-17 02:50:42 +08:00
rayd1o	9b913a3b83	release: bump version to 0.59.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-16 05:02:05 +08:00
linkong	93eb41a9f7	release: bump version to 0.58.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details Release 0.58.0 includes the Earth high-precision boundary PMTiles/MVT pipeline, standardized Earth boundary source collectors, China POV boundary configuration templates, and removal of the legacy low-precision GeoJSON fallback. It also adds Earth news target-location queueing/archive support, fixes datasource task status visibility, documents the Earth surface depth-spacing rules that prevent far-zoom z-fighting snow/black blocks, and updates bilingual operations/developer docs.	2026-05-15 17:40:07 +08:00
rayd1o	dd176a6ae6	release: bump version to 0.57.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details release / images (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details	2026-05-14 01:02:17 +08:00
linkong	f14ff6ec0f	release: bump version to 0.56.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-13 18:21:03 +08:00
linkong	39854b9983	release: bump version to 0.55.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-13 15:31:29 +08:00
linkong	3b4347c87d	release: bump version to 0.54.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-13 14:26:47 +08:00
rayd1o	d9efd98d26	release: bump version to 0.53.0 Some checks failed ci / backend (push) Has been cancelled Details ci / frontend (push) Has been cancelled Details ci / delivery (push) Has been cancelled Details release / images (push) Has been cancelled Details	2026-05-13 08:05:43 +08:00
@@ -1 +1 @@
 .52.0
 .60.0
				`@@ -0,0 +1,2 @@`
				`"""AI task prompt registry and runtime helpers."""`