codebase-summary

name	codebase-summary
description	Analyze any codebase — tree view, language breakdown, file counts by type, largest files, recent changes. Use for "summarize this project", "explore this codebase", "what's in this repo".
category	coding
version	1.0.0
examples	[{"prompt":"Summarize this repo and tell me how to run it."}]
test_commands	["python -m pytest tests/test_skills.py"]

Codebase Summary

Procedure

1. Tree the structure — Use file_tree or list_directory at the root to get the top-level layout. If the project is large, focus on src/, lib/, app/, packages/.

2. Count by language — Use run_command with language-appropriate counting:

   • find . -name '*.py' | wc -l and similar for .js, .ts, .rs, .go, .jsx, .tsx, .rs, .java, etc.
   • Include key config files: package.json, Cargo.toml, pyproject.toml, go.mod, etc.

3. Find largest files — run_command:

   • find . -type f -not -path './.git/*' -exec wc -l {} + | sort -rn | head -15

4. Check recent git activity if .git exists:

   • git log --oneline -10 for recent commits
   • git log --since="1 month ago" --format='%an' | sort | uniq -c | sort -rn for contributor activity

5. Write summary — Use write_file to save a markdown summary at the project root:

   • Project Overview

   • Structure (tree)

   • Language Breakdown

   • Largest Files

   • Command Inventory

   • Entrypoint Graph

   • Recent Activity (last 10 commits, contributors)

Architecture Map

• Identify app entrypoints, CLI scripts, server startup files, background workers, and test commands.
• List repeated command patterns from package scripts, pyproject metadata, Makefiles, CI, and README snippets.
• Sketch module relationships as text when a graphing tool is unnecessary.

6. Report — Tell the user the file path and key stats (total files, main language, most active areas).

Rules

• Do NOT read every source file — use tooling to count and size.
• Skip .git/, node_modules/, __pycache__/, .venv/, dist/, build/.