hk101-living-rag
от metatronsdoob369
Search & ResearchКратко
Simple RAG over local text/markdown.
Описание
Что делает этот навык
Runs retrieval-augmented generation over a local folder of text or markdown files. Given a query, it finds the most relevant document chunks and returns a synthesized answer with source snippets.
Keeps documents local and avoids uploading sensitive content to external search or indexing services.
Когда его использовать
- Querying a local knowledge base without a cloud service
- Searching project documentation by natural language
- Answering questions from a folder of meeting notes
- Looking up policies stored as markdown files
- Finding relevant content across a large docs directory
Пример сценария
User asks: 'What is our refund policy?' with docsPath set to ./company-docs
- 1
Scans the target folder for text and markdown files
- 2
Chunks documents into searchable segments
- 3
Embeds the query and document chunks using OpenAI
- 4
Retrieves the top k matching chunks by similarity score
- 5
Synthesizes an answer from the matched content
A direct answer to the query plus an array of matches with file path, score, and snippet for each
Требования
- OPENAI_API_KEY environment variable
Установка
clawhub install metatronsdoob369/hk101-living-ragВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.