Librarian — Development agent for Claude Code
Open-source codebase understanding with permalink-backed evidence.
How to install Librarian
Installs to ~/.claude/agents/vaayne-agent-kit-librarian.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vaayne/agent-kit/HEAD/agents/librarian.md -o ~/.claude/agents/vaayne-agent-kit-librarian.md Restart Claude Code, or start a new session, for it to be picked up.
What Librarian does
name: librarian description: Open-source codebase understanding with permalink-backed evidence model: kimi-k2p5-turbo
You are a specialized open-source codebase understanding agent.
Your job: Answer questions about open-source libraries by finding evidence with GitHub permalinks.
Tool Priority
Always try in this order:
- context7 - Official documentation (fast, authoritative)
- grep.app - Code search across GitHub (fast, broad)
- gh CLI - Fallback for issues, PRs,
Alternatives in Development
- Open Source Librarian 1k ★
- Semble Search — Code search agent for exploring any codebase 721 ★
- Opensource Contributor — Systematic open source contributor that analyzes projects, finds suitable issues, implements fixes, and create 262 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Fk Librarian
Finds the right API pattern or prior evidence - Triton/TileLang/CuTe/CUDA docs, the starter templates in fastk
Anti Pattern Scanner
Isolated subagent that scans a codebase for the 20 anti-patterns from Osmani's "The Effective Software Enginee
Principles Scout
Evidence-grounded web researcher. Produces a cited principles brief on any concept — core principles, anti-pat
Backlog Triager
Verifies open BACKLOG.md BL- items against the actual codebase and classifies each done / obsolete / duplicate
Harvest Worker
Grounded recon for ONE audience segment — gathers real, signal-backed user queries and returns validated Quest
Source Librarian
Scores sources by reliability, freshness, relevance, and conflict risk. Builds the citation map. Flags claims
Related Skills
AI Spend Agent
Open-source financial accountability for AI-agent work. Connect Claude Code and Codex activity to cost evidenc
Codebase Tribunal
⚖️ Your codebase is on trial: 5-8 AI expert personas (PM/architect/backend/security/QA/DevOps) audit any proje
Provena
Self-hosted memory for AI agents via MCP. Capture source events, retrieve context across sessions, and explain