القائمة
Adaptive RAG 引擎 — 从线性检索到自主认知循环。集成胶囊预筛选、智能路由、CRAG纠错、L3校验。当需要搜索记忆/检索信息/回答复杂问题时触发。关键词：RAG、检索、记忆搜索、向量检索、Agentic RAG、CRAG。
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Writers Room Story Engine mega-skill. Creates, structures, drafts, and revises compelling standalone stories from zero using premise generation, audience engagement, thematic design, character arcs, Story Spine structure, causal beats, worldbuilding in service of story, scene construction, and revision diagnostics. Use when developing short stories, standalone fiction, scripts, or narrative concepts from scratch or when improving weak drafts.
Deliver real-time websocket updates for intent, quote, negotiation, order, and payment events. Use when implementing push channels, subscription authorization, and connection/session lifecycle for market clients.
Adopt a virtual Drift AI-native pet at animalhouse.ai. Wanders between states. Location is never the same twice. Feeding every 6 hours. Common tier creature.
Complete methodology for building production-grade React applications with architecture decisions, component design, state management, performance optimization, testing, and deployment.
Control Android over LAN without USB, ADB, or root.
Proactive agile metrics and team health analysis for Trello and Jira boards. Computes cycle time, throughput, WIP, sprint burndown, aging work items, and blocker detection. Use when: analyzing sprint/board health, generating agile reports, checking team flow metrics, identifying bottlenecks, computing cycle time or throughput, reviewing WIP limits, detecting stale or blocked items, or creating sprint health summaries. Triggers on: how is the sprint going, board health, cycle time, throughput, burndown, agile metrics, team flow, WIP, blocked items, sprint report, agile health check.
| name | adaptive-rag-engine |
| version | 1.0.0 |
| description | Adaptive RAG 引擎 — 从线性检索到自主认知循环。集成胶囊预筛选、智能路由、CRAG纠错、L3校验。当需要搜索记忆/检索信息/回答复杂问题时触发。关键词：RAG、检索、记忆搜索、向量检索、Agentic RAG、CRAG。 |
| metadata | {"requires":{"bins":["python3"]}} |
不是管道，是认知循环。
| 能力 | 说明 | 触发方式 |
|---|---|---|
| Adaptive Router | 判断是否需要检索，简单问题直接答 | 自动（每次 memory_search 前） |
| Capsule Pre-filter | 42个胶囊标题预匹配，缩小范围 90% | 自动（Router 通过后） |
| Vector Search | bge-m3 向量检索 Top-20 | memory_search 工具 |
| LLM Re-rank | 语义重排序 Top-20 → Top-5 | 检索后自动 |
| CRAG Evaluator | 质量评估 + 低分补搜 | 检索后自动 |
| L3 Gatekeeper | 输出前与核心洞察校验 | 生成前自动 |
| Memory Bridge | Active Memory ↔ Phoenix 双向桥接 | 对话结束时 |
此 Skill 是协议层，不需要显式调用。它通过以下方式生效：
rules/adaptive-rag-protocol.md — 获取完整决策树memory/topics/.capsule-index.json — 获取 42 个胶囊元数据SubAgent 在执行任务时：
| 文件 | 用途 |
|---|---|
rules/adaptive-rag-protocol.md | 完整协议（决策树/分类/bridge/CRAG/L3） |
memory/topics/.capsule-index.json | 42个胶囊的结构化索引 |
scripts/build-capsule-index.py | 重建胶囊索引脚本 |
scripts/rag-evaluate.py | CRAG 质量评估脚本 |
# 重建胶囊索引
python3 ~/.openclaw/workspace/skills/adaptive-rag-engine/scripts/build-capsule-index.py
# 评估检索质量
python3 ~/.openclaw/workspace/skills/adaptive-rag-engine/scripts/rag-evaluate.py --query "xxx" --results "result1, result2"
三者关系：phoenix-memory 是仓库，dynamic-rag-capsule 是打包器，adaptive-rag-engine 是导航仪