Query wiki pages to answer codebase questions with cited references.
npx skills add https://github.com/baekenough/second-brain --skill wiki-rag-baekenough
Please help me install this Agent Skill. Skill: wiki-rag Source: https://github.com/baekenough/second-brain/tree/main/.claude/skills/wiki-rag Command: npx skills add https://github.com/baekenough/second-brain --skill wiki-rag-baekenough
Wiki-based codebase understanding is accelerated by leveraging a pre-compiled knowledge source rather than manually scanning raw files. This skill provides a fast, structured workflow to retrieve relevant wiki pages and synthesize concise, cited answers for codebase questions.
Ask a question about the codebase and let wiki-rag fetch, synthesize, and cite relevant wiki pages for an answer.
High-intent search queries and answers about installing and using this skill.
To search a project wiki for codebase questions using RAG, the skill loads wiki/index.yaml, identifies 3-7 relevant pages by title and category cues, and reads them in parallel to synthesize concise, cited answers with inline references.
The best way to answer codebase questions without scanning raw files is leveraging a pre-compiled wiki knowledge source. This approach retrieves relevant wiki pages via RAG, synthesizes direct cited answers, and evaluates coverage to determine if raw source exploration is still needed.
Yes, you need an index.yaml file to construct a catalog of pages. The skill specifically loads and parses wiki/index.yaml to identify relevant pages using title relevance, category cues, and linked relations before reading them in parallel.
Yes, you can automatically update wiki documentation when querying codebase answers. The skill delegates logging and wiki-update actions to subagents, maintaining a clean end-to-end workflow when updates are needed after synthesizing answers.
Wiki RAG coverage evaluation works by assessing whether the retrieved wiki pages fully answer the codebase question. After synthesizing cited answers from 3-7 relevant pages, it evaluates coverage to determine whether further raw source exploration is needed.
baekenoughBuilds and maintains an interlinked markdown wiki from project codebase.
soo-kate-yeonDiscover, load, and cache relevant documentation based on user intent and project context.
hiddink-aiQuery project wiki documentation to answer technical questions with inline citations.
ByronWilliamsCPATrace code dependencies and call chains using a pre-built knowledge graph.
