Braxis Keeps AI Agent Context Files in Sync With Codebases

Summary

Braxis is a Python developer tool for keeping the project context read by coding agents aligned with a codebase. Its `braxis generate` command analyzes a project and creates or updates AGENTS.md, CLAUDE.md, .cursorrules, and .agentic-config.json, so agents can use current structure and conventions. The project includes an agent-readiness score from 0 to 100, category-level breakdowns, and persistent score history with trend indicators. A GitHub Actions workflow can run scoring and regeneration after pushes and open a pull request when context files change; pre-commit hooks can also validate them. Braxis supports Python, JavaScript, TypeScript, Go, Rust, Java, and other languages, and its core package is described as having no external dependencies. Optional Claude API support provides five to seven implementation recommendations based on the project analysis, while the core commands work without an API key. The repository documents input validation, atomic file writes, path normalization, error handling, and more than 30 passing unit tests. It requires Python 3.8 or later, is released under the MIT license, and lists future plans including custom scoring rules, project comparisons, a web dashboard, and additional IDE integrations.