delegation
por michaelmonetized
Coding Agents & IDEsResumen
Architecture-first workflow for delegating complex projects to AI coding agents.
Descripción
Qué hace esta skill
Architecture-first workflow for delegating complex projects to AI coding agents. Every code change is analyzed against the system architecture before being written, with explicit filepath declarations, dependency mapping, and conflict checks built into the response format.
Forces AI agents to declare intent and validate fit before generating code, reducing the need to refactor or revert changes that violate existing architecture.
Cuándo usarla
- Adding a new API endpoint without breaking existing module contracts
- Onboarding an AI agent to a large unfamiliar codebase
- Generating production-ready code that matches existing naming conventions
- Catching duplicate functionality before it gets committed
- Ensuring new features stay within their intended architectural layer
Flujo de trabajo de ejemplo
User asks: Add a user preferences service to the backend
- 1
Reads architecture document to locate where a preferences service belongs
- 2
Declares target filepath and one-line purpose
- 3
Lists all imports the new service depends on and which modules will consume it
- 4
Checks for existing services that might duplicate this functionality
- 5
Outputs fully typed, production-ready code with error handling and matching test filepath
A preferences service file with explicit architecture analysis, typed implementation, test location, and a checklist confirming no linter, type, or architectural violations
Instalación
clawhub install michaelmonetized/delegationEjecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.