What problem does it solve? When multiple coding or analysis tasks can run independently, dispatching them to parallel agents without clear boundaries leads to merge conflicts, duplicated work, and inconsistent assumptions. This Skill provides a disciplined workflow for splitting work, scoping agents, and verifying combined results. ## Core Features & Use Cases - Independence Gating: Only dispatch in parallel when tasks have non-overlapping files, data, or decisions; otherwise run sequentially. - Scoped Dispatch: Each agent receives a narrow scope, acceptance criteria, forbidden areas, and required evidence to return. - Centralized Decisions: Shared architecture and API choices stay with the coordinator so parallel agents never make incompatible decisions. - Result Verification: Compare agent outputs against the plan, inspect changed files, and check for merge conflicts or duplicated work before continuing. - Use Case: During a deepwork session, split a refactor across three non-overlapping modules, dispatch one agent per module, then verify the combined diff for conflicts before merging. ## Quick Start Ask the agent to split these independent tasks across parallel agents with clear scopes and verify the combined result when they return.
