Maestro

If you're tired of Claude forgetting context between sessions or jumping straight to code without understanding requirements, this governance framework forces architectural discipline through a strict protocol. It mandates reading project memory files on startup, asks Socratic questions before executing, requires implementation plans for complex work, and routes tasks to specialized sub-skills for frontend, backend, or debugging. The TDD requirement and verification matrix are rigid, which is either exactly what you need for sprawling codebases or overkill for quick prototypes. The brain.jsonl memory system is clever for continuity, though the hooks are CLI-specific and won't fire in most IDE setups.

Install to Claude Code

npx -y skills add xenitv1/claude-code-maestro --skill maestro --agent claude-code