Andy Chan

End-to-end engineering harness for Claude Code — go from planning to building to reviewing in one workflow. 8 agents with multiple modes. In-built GUI to visualise your PRDs and todos. Ship with AI confidently. MIT.

When your agent starts coding, you gotta let it cook

Conventional Commits-style Skill for writing commits and PRs. Inbuilt changelog management. Github Workflows-compatible. <1K tokens per run to document commits accurately with breaking changes. MIT.

Writes the four files a repo needs before it goes public — README.md, LICENSE.md, SECURITY.md, llms.txt — from a scan of the repo itself. Claude Code skill, 8 modes, 7 protocols, 4 bundled figlet fonts. Never invents a fact: no guessed license, no fabricated benchmarks. MIT.

Rewrites your Claude Code prompts before the LLM sees them — detects code, question, or text-gen intent, strips noise, adds structure. Runs automatically via a UserPromptSubmit hook, no LLM call needed for most prompts. MIT.