What problem does it solve?
Breaks down complex goals into discrete, assignable sub-tasks and produces a clear execution plan so multiple specialized agents can collaborate without ambiguity or hidden dependencies. It prevents ad-hoc agent spawning by enforcing explicit skill assignments and dependency ordering, reducing coordination failures and duplicated effort.
Core Features & Use Cases
- Task decomposition: Split a high-level goal into independent or dependent subtasks suitable for parallel or sequential execution.
- Skill mapping: Assign each subtask to a specific named skill to ensure specialization and accountability.
- Dependency graph & aggregation: Mark task dependencies, indicate parallelizable phases, and specify how to combine outputs into the final deliverable.
- Use Case: Run a comprehensive pull request review by delegating code quality, security, and test coverage checks to different specialist skills and synthesizing the findings into a single verdict.
Quick Start
Decompose a pull request review into parallel tasks for code quality, security, and test coverage, assign each task a specific skill, define dependencies, and return a JSON execution plan.
