vlasenkoalexey

Wikify Repo — AI skill for Claude Code

AI community

Compile any codebase into a grounded, lint-clean markdown wiki your AI agent can actually trust — SCIP symbol resolution + citation linter, plain markdown you own (no graph DB, no SaaS).

How to install Wikify Repo

This entry records only its repository, not the path inside it, so there is no exact command to give. Open vlasenkoalexey/wikify-repo and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Wikify Repo does

Compile any codebase into a grounded, lint-clean markdown wiki your AI agent can actually trust — SCIP symbol resolution + citation linter, plain markdown you own (no graph DB, no SaaS).

Alternatives in AI

Jcodemunch MCP — Cut AI token costs 95%+ on code exploration 2.6k ★
Empirica — Make AI agents and AI workflows measurably reliable 244 ★
Scipilot Writing Skill — SciPilot family: academic writing and polishing co-pilot, with a machine-checked writing-quality evidence chai 220 ★

README

wikify-repo

**Compile any codebase into a knowledge-base wiki your AI agent can actually trust.**

![version](https://img.shields.io/badge/version-0.2.0-blue) ![python](https://img.shields.io/badge/python-3.11%2B-blue) ![license](https://img.shields.io/badge/license-MIT-green) ![works with](https://img.shields.io/badge/works%20with-Claude%20Code%20%C2%B7%20Codex%20%C2%B7%20Antigravity-8A2BE2)

**wikify-repo** turns a repository into a grounded, lint-clean [Karpathy-style LLM markdown wiki](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f): every claim on every page cites a real, compiler-resolved symbol, and a citation linter fails the build if one doesn't check out. No graph database, no dashboard, no hosted service — the output is plain markdown in your own git repo, and your agent answers from it with nothing but `grep`. A deterministic tool does the grounding (SCIP symbol graph, packets, citation lint, coverage); one LLM-in-the-loop step writes the prose. Record every class, method and relationship with SCIP, spend the model on the ~20% of the code that explains ~80% of it, and give the rest a deterministic catalog page so nothing is dropped.

**For:** teams running coding agents on a codebase too large to read; knowledge bases that span many repositories; onboarding without a tour guide.

[![Force-directed graph of a wiki produced by wikify-repo: two ingested codebases plus prose pages, colored by page type — click for the interactive view](assets/demo-graph.png)](https://vlasenkoalexey.github.io/wikify-repo-demo/tools/graph/) A wiki produced by this tool — two ingested codebases plus prose pages, every link a citation. Click for the [interactive view](https://vlasenkoalexey.github.io/wikify-repo-demo/tools/graph/).

30 seconds

pipx install git+https://github.com/vlasenkoalexey/wikify-repo && wikify setup   # CLI + agent skill
cd my-repo && wikify init                # wiki config + a block in CLAUDE.md / AGENTS.md