Auto-generate AI agent context files. Keep them in sync with your code.
Your AI agents (Claude Code, Cursor, Copilot) read from AGENTS.md to understand your project. When your code changes, that file gets stale. Agents miss patterns, violate conventions, hallucinate.
Braxis solves this: one command generates four context files that stay in sync with your codebase.
Before Braxis:
Day 1: Agent reads stale AGENTS.md from 2 weeks ago Sees old directory structure Doesn't know about new error handling pattern Makes bad suggestions based on outdated info
After Braxis:
Every push: GitHub Actions runs Braxis Analyzes current codebase Regenerates AGENTS.md, CLAUDE.md, .cursorrules, .agentic-config.json Creates PR with updates Your agents always see current reality
- ✅ One Command - Generate all context files with braxis generate
- ✅ Zero Config - Works out of the box, no setup needed
- ✅ Auto-Score - Measure your project's AI agent readiness (0-100)
- ✅ Score History - Track improvements over time with trends & analytics
- ✅ LLM Recommendations - AI-powered suggestions using Claude API (optional)
- ✅ Multi-Language - Supports Python, JavaScript, TypeScript, Go, Rust, Java, and more
- ✅ CI/CD Ready - GitHub Actions workflow included
- ✅ Pre-commit Hooks - Validate before every commit
- ✅ Safe & Reliable - Input validation, atomic writes, comprehensive error handling
- ✅ Well-Tested - 30+ unit tests with 100% pass rate
- ✅ No Dependencies - Pure Python, zero external packages (LLM features optional)
- ✅ Production-Grade - Used in real projects, actively maintained
# Basic installation (core features)
pip install braxis
# With LLM support (for AI recommendations)
pip install braxis[llm]For LLM-powered recommendations:
export ANTHROPIC_API_KEY='sk-ant-...'Get your API key: https://console.anthropic.com
cd /path/to/your/projectbraxis generateThat's it. Braxis creates:
your-project/ ├── AGENTS.md (Universal agent instructions) ├── CLAUDE.md (Claude Code optimized) ├── .cursorrules (Cursor IDE rules) ├── .agentic-config.json (Machine-readable metadata) └── (your existing files)
cat AGENTS.mdgit add AGENTS.md CLAUDE.md .cursorrules .agentic-config.json
git commit -m "chore: add AI agent context files"
git pushIn Claude Code: Automatically reads CLAUDE.md
In Cursor: Copy .cursorrules into Cursor Settings → Rules
In any agent: Reads AGENTS.md (universal format)
Automatically regenerate context files on every push using GitHub Actions.
Braxis includes a ready-to-use workflow. Copy it to your repo:
mkdir -p .github/workflows
cp /path/to/braxis/.github/workflows/braxis-score.yml .github/workflows/
git add .github/workflows/braxis-score.yml
git commit -m "chore: add braxis auto-update workflow"
git pushOr manually create .github/workflows/braxis-score.yml:
name: Braxis Score Check
on:
push:
branches: [ main, develop ]
paths:
- '**.py'
- 'package.json'
- 'pyproject.toml'
- 'setup.py'
jobs:
score:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v4
with:
python-version: '3.11'
- run: pip install braxis
- run: braxis score
- run: braxis generate
- name: Create Pull Request for updates
uses: peter-evans/create-pull-request@v5
with:
commit-message: 'chore: regenerate braxis context files'
title: 'chore: update agent context files'
branch: braxis/auto-updateResult: Every push automatically regenerates context files and creates a PR if needed. ✨
Validate context files before every commit using pre-commit.
pip install pre-commit
pre-commit installThe hooks will run automatically on git commit.
Copy the example config to your project:
cp /path/to/braxis/.pre-commit-config.example.yaml .pre-commit-config.yamlThen install:
pip install pre-commit
pre-commit installNow braxis will validate your project before each commit! 🔐
Track your project's AI agent readiness score over time.
# View all historical scores
braxis history
# View scores with trends and direction indicators
braxis history --trendsExample output:
============================================================
Score History for myproject
============================================================
1. 2026-10-01 - 65/100 (AI-Native)
2. 2026-10-05 - 72/100 (AI-Native)
3. 2026-10-10 - 78/100 (AI-Native-Plus)
Trend: 📈 +13 points
============================================================
Features:
- Automatic score persistence on every braxis scorerun
- Project-specific tracking (stored in ~/.braxis/history/)
- Trend indicators: 📈 (improving) 📉 (declining) ➡️ (stable)
- Configurable history limits
- Timestamps and tier information included
Get intelligent, actionable recommendations from Claude AI.
Setup:
# Install with LLM support
pip install braxis[llm]
# Set your API key
export ANTHROPIC_API_KEY='sk-ant-...'Usage:
braxis recommendationsExample output:
============================================================
LLM-Powered Recommendations for myproject
============================================================
1. Add Comprehensive Test Suite
Why it matters: Testing is the foundation of reliable code.
Current state: Only 7% testing coverage
How to implement:
- Start with pytest fixtures for common patterns
- Aim for 80%+ coverage on core modules
- Run: pytest --cov to measure progress
2. Implement Input Validation Framework
Why it matters: Validation prevents bugs and security issues
Current state: No systematic validation detected
How to implement:
- Use Pydantic for request validation
- Add schema validation to all API endpoints
- Example: from pydantic import BaseModel
[... more recommendations ...]
Features:
- Uses Claude Opus 5.5 for high-quality analysis
- 5-7 actionable recommendations per run
- Concrete implementation steps for each suggestion
- Focuses on improving Agent Readiness Score
- Gracefully handles missing API keys
- Optional dependency (braxis works without it)
Braxis welcomes contributions! Read CONTRIBUTING.md for:
- Setup instructions
- Development workflow
- Testing guidelines
- Code style guide
- PR process
Quick start:
git clone https://github.com/YOUR_USERNAME/braxis.git
cd braxis
python3 -m venv venv
source venv/bin/activate
pip install -e .
python3 -m unittest test_braxis -vCreate .agentic-config.json:
{
"name": "MyApp",
"description": "A production API service",
"exclude_patterns": [
"node_modules/**",
".venv/**",
"build/**"
],
"custom_conventions": {
"error_handling": "Always use try/except and log",
"async_patterns": "All I/O must be async",
"validation": "Use Pydantic models for inputs"
},
"critical_files": [
"src/main.py",
"src/api/routes.py",
"README.md"
]
}AGENTS.md - Universal format read by any AI agent
CLAUDE.md - Optimized for Claude Code
.cursorrules - Rules for Cursor IDE
.agentic-config.json - Machine-readable metadata
Check how ready your codebase is for AI agents:
braxis scoreExample output:
============================================================
Agent Readiness Score: 71/100
============================================================
Breakdown:
Architecture 10/100 [██░░░░░░░░░░░░░░░░░░]
Testing 7/100 [█░░░░░░░░░░░░░░░░░░]
Dependencies 12/100 [██░░░░░░░░░░░░░░░░░]
Conventions 10/100 [██░░░░░░░░░░░░░░░░░░]
Entry Points 4/100 [░░░░░░░░░░░░░░░░░░░]
Security 10/100 [██░░░░░░░░░░░░░░░░░░]
Build 10/100 [██░░░░░░░░░░░░░░░░░░]
Documentation 8/100 [█░░░░░░░░░░░░░░░░░░]
Tier: AI-Native
Detected:
Languages: python
Build System: Python (pip/setuptools)
Test Frameworks: pytest, unittest
Test Files: 1
Critical Files: 0
Recommendations:
* Increase test coverage
* Add input validation and security checks
- Architecture - Critical files, entry points, project structure
- Testing - Test coverage and test framework detection
- Dependencies - Build system and dependency management
- Conventions - Code patterns, error handling, type hints
- Entry Points - Main functions and executable files
- Security - Input validation, security checks, config management
- Build - Build files and dependency tracking
- Documentation - README and project documentation
python3 braxis.py score --path /path/to/project# Check version
braxis --version
# Score your project's agent readiness
braxis score
braxis score --path /path/to/project # Score a specific project
# Generate context files
braxis generate
braxis generate --path /path/to/project
# Inspect project analysis
braxis inspect
braxis inspect --path /path/to/project
# Validate context files exist
braxis validate
braxis validate --path /path/to/project
# View score history
braxis history
braxis history --path /path/to/project
braxis history --trends # Show trends with emoji indicators
braxis history --path /path/to/project --trends
# Get LLM-powered recommendations (requires: export ANTHROPIC_API_KEY='sk-ant-...')
braxis recommendations
braxis recommendations --path /path/to/projectBraxis is production-grade with enterprise-level quality standards:
- Input Validation - Validates project paths and file inputs with clear error messages
- Atomic File Writing - Uses temporary files and atomic operations to prevent partial writes
- Error Handling - Comprehensive error handling with informative feedback
- Path Normalization - Converts relative paths to absolute paths safely
- Comprehensive Tests - 30+ unit tests covering all major functionality
- Test Coverage - 100% pass rate across all test suites
- Automated Testing - GitHub Actions runs tests on every commit
- Pre-commit Hooks - Validates before every commit
- Code Style - Follows PEP 8 standards
- Zero Dependencies - No external packages required
# Run all tests
python3 -m unittest test_braxis -v
# Run specific test class
python3 -m unittest test_braxis.TestValidateProjectPath -v
# Check test coverage
pip install coverage
coverage run -m unittest test_braxis
coverage reportStatus: All 30 tests passing ✅
The difference: Braxis goes beyond rules files. It continuously analyzes your codebase, scores your readiness, tracks progress, and provides AI-driven guidance—all automatically.
- Python 3.8+
- Zero external dependencies
- Works on macOS, Linux, Windows
- Report Issues - GitHub Issues
- Discussions - GitHub Discussions
- Contributing - See CONTRIBUTING.md
# New developer clones repo
$ braxis score
Agent Readiness: 78/100 (AI-Native-Plus)
# They immediately understand the project structure, conventions, and quality baseline
# Opens AGENTS.md in Claude Code - instant context# Initial state
$ braxis score
Score: 45/100 (Agent-Aware)
# After 2 weeks of improvements
$ braxis history --trends
📈 +28 points
# Team celebrates progress with visual trend data# Every commit triggers workflow
$ braxis score
$ braxis generate
# PR created with updated context files
# Agents always have latest project info$ braxis recommendations
# Get AI-powered guidance on what to improve
# Prioritize high-impact changes
# Measure progress with `braxis history`AI agents need current context to work effectively. Without it, they:
- ❌ Miss recent code patterns
- ❌ Violate project conventions
- ❌ Make outdated suggestions
- ❌ Waste your time with hallucinations
Braxis solves this automatically:
- Generates context files from real code analysis
- Scores your readiness for AI agents (0-100)
- Tracks improvements over time with trends
- Recommends actionable next steps via Claude AI
One command. Always in sync. Always improving. ✨
MIT - Free to use in personal and commercial projects
See LICENSE for details.
Upcoming features in development:
- 🚧 Custom scoring rules engine
- 🚧 Project comparison & benchmarking
- 🚧 Web dashboard for visualization
- 🚧 Score forecasting & predictions
- 🚧 Integration with more IDE platforms
Have a feature request? Open an issue
Braxis — Keep your agents aligned. Keep your code context current.
Continuously analyze. Automatically improve. Always sync. ✨
Made with ❤️ for AI-native development by developers, for developers.