Added comprehensive safety features to prevent accidental commits and ensure
user review of all generated commit messages:
Safety enhancements:
- Added mandatory review warning with clear visual indicators
- Implemented double confirmation for all commit operations
- Added option to regenerate commit message if not satisfied
- Shows edited message for review after modifications
- Added explicit "no auto-push" reminders after commits
- Created ~/.claude-commit.conf for safety preferences
Documentation updates:
- Updated CLAUDE.md with clear commit workflow for Claude Code
- Added safety features section to tool documentation
- Clarified that manual review is always required
- Documented the review-approve-commit-push workflow
These changes ensure users always have full control over commit messages
and prevent accidental commits or pushes without explicit approval.
Implemented a comprehensive AI-powered commit message generator using Claude that:
- Automatically generates Conventional Commit formatted messages
- Analyzes git diff to create context-aware commit messages
- Works globally across all git repositories with 'git cc' command
- Supports multi-computer setup through portable dotfiles
Major changes:
- Added global claude-commit script with git aliases (git cc, git smart-commit)
- Created organized documentation in docs/tools/claude-commit/
- Included portable dotfiles structure for easy multi-machine deployment
- Updated README with TLDR Node.js quickstart section featuring all Brainy capabilities
- Moved documentation section to bottom of README for better flow
- Added CLAUDE.md with project-specific instructions for Claude Code
The tool eliminates manual commit message writing by leveraging AI to understand
code changes and generate properly formatted, meaningful commit messages that
follow the Conventional Commits specification.