dispatching-parallel-agents

Debug multiple unrelated test failures by assigning each independent failure domain to a separate, focused agent. This parallel delegation prevents wasted time by allowing specialized agents to investigate distinct problems simultaneously.

Use parallel agents for task delegation when you have independent technical tasks without shared state or sequential dependencies. This includes unrelated test files, subsystem investigations, and debugging tasks that can be processed simultaneously.

Coordinate independent debugging workstreams by providing each agent with self-contained prompts, explicit constraints, and expected outcomes. After execution, review the results, check for conflicts, and perform final verification together.

Yes, you can delegate unrelated subsystem investigations to focused agents. Ensure each investigation is truly independent and lacks sequential dependencies so agents can operate in parallel without conflicting state modifications.

Dispatching parallel agents is limited to tasks without shared state or sequential dependencies. It cannot be used for workstreams that require sequential execution, and it requires careful conflict checking and full-suite testing to verify combined results.

The best way to verify fixes from parallel debugging agents is to integrate the results, check for conflicts between workstreams, run full-suite testing, and perform spot verification to ensure the combined fixes are stable.