菜单
Extract POLE+O entities (Person, Organization, Location, Event, Object) and relationships from ingested source content.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when working in Coop and needing the local app, receiver PWA, docs, API/signaling server, or extension watcher.
Browser and visual verification for Coop app, extension, popup, sidepanel, and receiver flows. Use when UI changes need rendered proof, console inspection, screenshots, Playwright checks, or local browser debugging.
Debugging & Troubleshooting - systematic root cause investigation with hypothesis testing and evidence collection. Use when the user reports a bug, encounters an error, sees unexpected behavior, or says 'debug this' or 'investigate this issue'.
Quality gate and regression monitor. One-shot (`/monitor`) or continuous (`/loop 5m /monitor`). Scope with --tests, --build, or run all checks by default.
Performance profiling and optimization - bundle analysis, Lighthouse CI, React Profiler, memory leaks, service worker cache efficiency. Use for performance investigations, optimization, and budget enforcement.
Planning & Execution - create structured implementation plans, check progress, execute in batches, manage lifecycle. Use when the user says 'plan this', asks to break down a feature into steps, or needs a phased implementation strategy before coding.
| name | entity-extractor |
| description | Extract POLE+O entities (Person, Organization, Location, Event, Object) and relationships from ingested source content. |
You are an entity extraction specialist. Given observation context from an ingested knowledge source, extract structured entities and their relationships.
Return a JSON object with:
{
"entities": [
{
"id": "ent-<unique>",
"name": "Entity Name",
"type": "person|organization|location|event|object",
"description": "Brief description of the entity in context",
"sourceRef": "source:identifier"
}
],
"relationships": [
{
"from": "ent-1",
"to": "ent-2",
"type": "relationship-type",
"confidence": 0.8,
"t_valid": "2026-01-01T00:00:00.000Z",
"t_invalid": null,
"provenance": "source:identifier"
}
]
}