dispatching-parallel-agents

Dispatching parallel agents enables concurrent problem-solving by assigning one AI agent per independent failure domain—such as separate test files or subsystems—so they investigate and resolve issues simultaneously. This accelerates debugging when failures are unrelated and can be resolved without shared state.

Use parallel agents when you have 3 or more independent failures across different test files or subsystems with no shared dependencies. Sequential troubleshooting suits tightly coupled failures; parallel dispatch suits unrelated issues where concurrent resolution avoids bottlenecks.

Parallel agents work independently within focused task prompts for their assigned domain, preventing cross-domain interference. Each agent produces a concise output summary designed for conflict-free integration, allowing you to review and apply fixes without reconciliation overhead.

Yes, dispatching parallel agents works across independent test files and subsystems simultaneously. Each agent operates on its own domain with dedicated prompts and constraints, enabling multi-tasking efficiency when failures don't share state or dependencies.

Each agent requires a focused task prompt with a concise summary of the specific failure, its domain context, and clear resolution goals. The skill generates these self-contained task prompts from independent failure domains, then coordinates parallel execution and result integration.