knowledge-retriever

name	knowledge-retriever
description	Retrieves relevant knowledge from project documentation, code comments, and README files
version	1.0.0
author	go-on-team
tags	["knowledge","retrieval","documentation","search","context"]
min_go_on_version	1.0.0

Knowledge Retriever Skill

Scans project documentation, comments, READMEs, and design docs to surface the most relevant information for a given query. Useful for onboarding, answering architecture questions, or gathering context before making changes.

How It Works

1. Index — The skill crawls known documentation files (README.md, CONTRIBUTING.md, ARCHITECTURE.md, CHANGELOG.md, doc-comments in source, and any docs/ directory) and builds a lightweight search index.
2. Query — Your natural-language query is matched against the index using keyword and semantic similarity.
3. Rank — Results are scored by relevance and returned up to the requested max_results.
4. Present — Each result includes the source file path, a snippet of the surrounding context, and a relevance score.

Input Schema

Parameter	Type	Description
query	string	Natural-language question or keywords to search for
max_results	integer	Maximum number of results to return (default: 5, max: 20)

Example

{
  "query": "How does the authentication middleware work?",
  "max_results": 3
}

Example Output

{
  "results": [
    {
      "file": "docs/auth/overview.md",
      "snippet": "The JWT middleware is registered in src/main.rs on line 42. It extracts the Bearer token from the Authorization header...",
      "score": 0.94
    },
    {
      "file": "src/auth/middleware.rs",
      "snippet": "// validate_token decodes the JWT, checks expiry, and injects the Claims into the request context.",
      "score": 0.87
    },
    {
      "file": "README.md",
      "snippet": "Authentication: set AUTH_SECRET env var and include an Authorization: Bearer <token> header on protected routes.",
      "score": 0.72
    }
  ]
}