One command to know your codebase health. 31 CLI commands, 8 analysis dimensions, MCP server for AI agents. MIT licensed. Local-first. Zero config.
pip install chm-cli
chm init .
chm doctor .Chm reads your git history and gives you a health report — not opinions, but data:
| Dimension | Tells you |
|---|---|
| 🔥 Hotspots | Which files change too often (bug factories) |
| 👥 Contributors | Bus factor — how fragile is your team? |
| 💓 Team Pulse | When does your team actually work? |
| 🧩 Complexity | Hard-to-maintain files with low comments |
| 💀 Dead Code | Files untouched for months — candidates for deletion |
| 🔗 Dependencies | Circular deps and over-coupled modules |
| 🧪 Test Coverage | Which files lack tests (by naming convention) |
| 📋 Duplication | Copy-pasted code across files |
$ chm doctor .
🏥 CHM Doctor — Diagnosis for my-api
Health Score: 62/100
🚨 [Team Risk] Bus factor = 1 — single point of failure
▸ 立即开始结对编程,让第二个人熟悉核心模块
▸ 核心文件:src/auth.py, src/payment.py
⚠️ [Hotspots] Top 3 files account for 52% of churn
▸ 优先重构 src/payment.py — 被 3 个人改了 12 次
▸ 为这些文件补充单元测试# Install
pip install chm-cli
# Try it without a real project
chm demo
# Analyze your repo
chm init . # First-time wizard
chm analyze . # Full 8-dimension report
chm doctor . # Actionable prescriptions
# Generate shareable report
chm analyze . --report html -o health.html
# Track trends over time
chm snapshot . # Save a baseline
chm trends . # See how health changes
# Generate a README badge
chm badge . -o badge.svgChm is a Model Context Protocol server — AI agents can query your codebase health directly:
# Register with Claude Code
claude mcp add chm -- chm mcp
# Then ask Claude:
# "What are my top hotspots?"
# "What's my bus factor?"
# "Which files need tests the most?"12 MCP tools + 1 resource exposed.
pip install chm-cli # Base CLI
pip install chm-cli[mcp] # With MCP server supportRequires Python 3.9+ and Git.
| Command | Description |
|---|---|
analyze |
Full 8-dimension health analysis |
doctor |
Diagnosis with prioritized prescriptions |
init |
Interactive onboarding wizard |
demo |
Run demo without a real repo |
hotspots |
Most-changed files |
authors |
Contributor stats + bus factor |
pulse |
Team activity rhythm |
churn |
Code churn overview |
complexity |
Code complexity analysis |
deadcode |
Stale file detection |
deps |
Dependency + coupling analysis |
coverage |
Test coverage estimation |
duplicates |
Code duplication detection |
snapshot |
Save health baseline |
trends |
View health trends |
compare |
Compare two snapshots |
badge |
Generate SVG badge |
watch |
Auto-analyze on git commit |
mcp |
Start MCP server |
login |
Connect SaaS account |
status |
SaaS entitlement check |
An optional SaaS backend provides:
- 📊 Multi-repo team dashboards
- 📧 Weekly email reports
- 🔑 License key management
- 💳 Subscription billing (Stripe/PayPal/WeChat/Alipay)
| CHM | SonarQube | CodeClimate | GitCompass | |
|---|---|---|---|---|
| Install | pip install |
Java server | SaaS signup | npm install |
| Price | Free / $29 | $150+ | $49+ | Free |
| Local-first | ✅ | ❌ | ❌ | ✅ |
| Open source | ✅ MIT | ❌ | ❌ | ✅ MIT |
| MCP support | ✅ 12 tools | ❌ | ❌ | ❌ |
| SaaS backend | ✅ | ❌ | ✅ | ❌ |
MIT — free for individuals and companies. SaaS platform is proprietary.
Built with ❤️ by Lighthouse Analytics