workflow-designer

Design AI coding workflows by analyzing task requirements against a three-layer model of agents and skills. This framework evaluates interactions between Main Agents, Subagents, and Skills to orchestrate processes and prevent context rot.

You need a subagent for isolated code exploration tasks and a reusable skill for standardized actions like PR submission. Evaluating tasks against a three-layer architecture determines the correct approach and prevents over-engineering.

The three-layer architecture model structures AI development through the interaction of Main Agents, Subagents, and Skills. It provides a rigorous framework to decompose tasks, orchestrate subagents, and apply design principles for scalable environments.

Prevent context rot by applying core design principles like Information Isolation and Read-Before-Write. Enforcing these principles during workflow architecture ensures safe, efficient, and noise-isolated task execution across agents.

Yes, the workflow architecture explicitly supports task decomposition for large coding projects. It analyzes task requirements to orchestrate subagents and define reusable skills, satisfying the need for consistent and scalable development environments.