菜单
Full Qdrant data reset — run when embedding dimension changes or data is corrupted. Destroys all vector data.
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
Create test cases, run QA/QC checks, and ensure code quality. Use when asked to write tests, review code quality, or perform QA. Supports: unit tests, integration tests, coverage analysis, linting, type checking.
Diagnose why the agent isn't calling wiki tools correctly — checks tool registration, system_instruction wiki strategy, and replays a query to trace tool call behavior.
Standard flow for adding a new research paper to MemRAG — upload, wait for wiki ingestion, verify output. Use when onboarding a new paper into the knowledge base.
Inspect wiki state for a user — list all pages, read index.md, show recent log entries. Use when wiki isn't updating correctly after ingestion or when you need to verify wiki content.
Diagnose và fix các vấn đề liên quan đến Google ADK agents, tools, session state, multi-agent communication trong dự án MemRAG.
Run the full pre-push verification sequence — backend lint+format, backend tests, frontend types+lint+build. Stop and report on first failure.
| name | docker-reset |
| description | Full Qdrant data reset — run when embedding dimension changes or data is corrupted. Destroys all vector data. |
| disable-model-invocation | true |
This destroys all Qdrant vector data (RAG documents and mem0 memories). Use when:
Run from the project root (/home/minhdd/pet_proj/proj2):
docker compose down -v && docker compose up -d
Then wait for Qdrant to be healthy before restarting the backend. Collections will be auto-created on next app startup via ensure_collections().