What problem does it solve? Coordinating AI agents on real codebases often leads to blurred responsibilities: agents plan and implement at once, review their own work, or commit without authorization. This Skill enforces explicit role separation, bounded Git authority, and traceable handoffs so development work stays supervised and auditable. ## Core Features & Use Cases - Role Routing: Directs each request to exactly one role—analyst, developer, debugger, coding-agent, fix-agent, or code-reviewer—based on request signals, with a fast Retouch Mode for trivial localized changes. - Architecture Gate Enforcement: Blocks analysis and implementation until a required architecture gate in the work contract reports Status: passed, preserving the Architect Addendum. - Controlled Git Authority: Standalone roles never commit or merge without direct human authorization; in orchestrated mode the project-orchestrator owns all Git lifecycle operations. - Use Case: A user asks to fix a regression in a Git repository. The Skill routes to the debugger role, which reproduces the failure, identifies the root cause, implements a minimal fix with regression tests, and hands off uncommitted work to a distinct code-reviewer with a suggested commit message. ## Quick Start Ask the agent to implement a feature or fix a bug in your repository and it will select the appropriate development role, follow the gated workflow, and hand off reviewed work with a suggested commit message.
