Build a local index and search the PAI knowledge base with weighted keyword matching.
npx skills add https://github.com/matt-fell-ai/PAI --skill librarian
Please help me install this Agent Skill. Skill: librarian Source: https://github.com/matt-fell-ai/PAI/tree/main/skills/Librarian Command: npx skills add https://github.com/matt-fell-ai/PAI --skill librarian
Librarian provides deep search and retrieval across the PAI knowledge base, helping you locate past solutions, related code, and project context quickly.
High-intent search queries and answers about installing and using this skill.
Search your knowledge base by running indexed queries across History, Skills, and Memory. Librarian builds a local index using keyword matching and returns the top 5 most relevant results with relevance scores, letting you quickly locate past decisions, code patterns, and project context without manual browsing.
Yes. Librarian performs weighted keyword matching across your indexed knowledge base to surface conceptually related code, prior implementations, and linked decisions. It filters noise from system directories and ranks results by relevance, making it easy to discover reusable patterns and avoid duplicating effort.
Librarian builds a local index using ripgrep if available for speed, otherwise falls back to Python's os.walk for filesystem traversal. It applies weighted keyword matching against indexed content, filters common noise directories, and returns ranked results so retrieval is both fast and accurate.
Yes. Librarian indexes across History, Skills, and Memory domains simultaneously, handling searches that span multiple knowledge sources. The weighted ranking system scales to large histories and returns the top 5 results, so you get the most relevant matches even in complex, multi-domain projects.
Yes. Search by project name, topic, or problem type to retrieve relevant background information and prior solutions. Librarian surfaces linked decisions and contextual details from your knowledge base, making it ideal for quick onboarding, debugging, and understanding why past choices were made.
thecloudtipsCrawl Notion and Google Drive to catalog knowledge sources with metadata and classification.
hoanghn61Converts TechStackReport JSON into Markdown, HTML, and PDF exports.
sakamoto-family-smileStore and organize knowledge across local files, MCP memory, vector stores, and Git repositories.
as4584Builds Retrieval-Augmented Generation systems for LLM applications using vector databases and semantic search.
