What problem does it solve? Coordinating multiple parallel AI workers manually leads to inconsistent briefs, lost results, and unclear completion criteria. This Skill structures parallel fan-out so every worker gets a standalone brief, results are aggregated against a declared done predicate, and gaps or dropouts are reported honestly instead of guessed. ## Core Features & Use Cases - Structured Fan-Out: Spawn N cloud workers in one message with standalone briefs covering goal, scope, verification, and reporting format. - Three Shapes: Partition work into slices, race N workers on identical briefs with a declared selection rule (first pass, rank all, best-of), or mix both. - HOLD Discipline: Stop and report missing inputs (done predicate, worker count, output separation) rather than inventing assumptions. - Use Case: Ask to swarm a codebase audit across five modules; each worker covers one slice, reports PASS/ISSUES/BLOCKED with evidence, and you receive one compact table with issues and explicit gaps. ## Quick Start Ask the agent to swarm this task with four parallel workers, each covering a separate module, and return one consolidated report.
