harness-configs
♥Cherished
Configuration templates for Claude Code and Codex that implement structured AI agent workflows with enforced permission models, automated code quality checks, and git branching isolation. Built for developers who want to systematize how their AI coding assistants handle permissions, formatting, linting, and code review—without manual intervention.
Clauded With Love Rating
8.1 / 10
This project provides sophisticated configuration templates for Claude Code and Codex that implement structured AI agent workflows with automated quality checks, git branching isolation, and granular permission models. It systematizes AI coding assistant behavior through comprehensive rules, automated formatting/linting, and specialized commands for code review and git operations.
Code Quality6.5
Usefulness8.2
Claude Usage9.1
Documentation7.8
Originality8.7
Highlights
- ✓Innovative granular bash permission system that auto-allows read-only commands while explicitly requiring approval for write operations like git commits
- ✓Comprehensive two-pass code review system with automated mechanical fixes and intelligent scope drift detection
- ✓Sophisticated git worktree workflow that automatically isolates branches across sessions for non-trivial tasks
To Improve
- →Add error handling and validation to the setup script, particularly for missing dependencies and failed symlink operations
- →Include example configuration files and usage scenarios to help users understand the practical impact of different settings