dispatching-parallel-agents

Debug multiple independent failures concurrently by dispatching parallel AI agents, assigning one agent per isolated problem domain to investigate root causes simultaneously and return structured summaries.

Parallel agent dispatch is a task orchestration technique that assigns one agent per independent failure domain to work concurrently without shared state, producing isolated diagnostics for review.

Dispatch parallel agents for debugging when you have 3 or more independent issues, such as failing tests or separate subsystems, where each problem can be understood without context from the others.

Start a parallel investigation by outlining 3 or more independent failures and requesting parallel investigation; the skill dispatches agents to explore each domain and return concise diagnostics and next steps.

Parallel agents are not suited for dependent subsystems with shared state. The skill ensures domain isolation to avoid cross-contamination, requiring each investigation to be understood independently without context from others.

Dispatching parallel agents works with separate failing tests by assigning one agent per failing test domain to work concurrently, ensuring isolated scopes and returning structured output summaries for comparison.