learn-repo

Analyzes a repository's code patterns and writes structured markdown learnings to a local directory.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/muselesscreator/ui-skills --skill learn-repo-muselesscreator
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: learn-repo
Source: https://github.com/muselesscreator/ui-skills/tree/main/learn-repo
Command: npx skills add https://github.com/muselesscreator/ui-skills --skill learn-repo-muselesscreator

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Starting work on an unfamiliar codebase requires hours of manual exploration to understand its component patterns, conventions, test structure, and tooling rules. This Skill automates that discovery by analyzing the repository's source code and persisting the findings as reusable markdown files. ## Core Features & Use Cases - Full or Scoped Analysis: Analyze an entire repository or scope the analysis to a specific path or component via an optional argument. - Pattern Extraction: Captures UI component patterns, data-layer conventions, file organization, naming conventions, test patterns (Vitest/Cypress), and tooling enforcement (TypeScript strict flags, ESLint rules, Prettier settings). - Structured Output: Writes flat markdown files (index.md, ui-patterns.md, file-structure.md, gotchas.md, test-patterns.md, standards.md) to ~/.claude/repo-learnings/{repo}/ for later reuse. - Use Case: Before contributing to an unfamiliar open-source project, run the analysis once to get a synthesized reference of its conventions, reference implementations, and known gotchas instead of reading every file manually. ## Quick Start Ask the assistant to learn this repository's patterns and conventions before you start working on it.

Frequently Asked Questions about learn-repo

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze an unfamiliar codebase before contributing?

Run the learn-repo analysis on the repository root to extract component patterns, naming conventions, test structure, and tooling rules. The results are written as markdown files to ~/.claude/repo-learnings/{repo}/ for later reference.

How to analyze only one feature or folder in a repo?

Pass a path as an argument to scope the analysis to that directory. Scoped learnings are written to ~/.claude/repo-learnings/{repo}/scoped/{slug}.md with a pointer added to index.md.

Does repo analysis work on projects that have a wiki directory?

No. If the repository contains a ./wiki/ directory, the analysis stops and routes you to the wiki-braindump and wiki-ingest workflow instead, since canonical knowledge for work repos belongs in the wiki.

What happens if learnings already exist for a repository?

The analysis short-circuits when an existing index.md is found in the repo's learnings directory. You are directed to use update-learnings to refresh, unless you explicitly request a full re-analysis.

What information does the codebase analysis capture?

It captures UI component and data-layer patterns, file organization and naming conventions, TypeScript strict flags, ESLint and Prettier rules, unit and E2E test patterns, code gotchas from NOTE/HACK/TODO comments, and 1-3 reference implementations.