dispatching-parallel-agents

Parallel debugging dispatches one focused agent per independent test file or subsystem to investigate and fix issues concurrently, saving time and context switching. You collect summaries from each agent and run the full test suite before integration.

Fixing unrelated CI test failures is best handled by dispatching parallel agents, where each agent receives a narrow scope and self-contained prompt to diagnose a separate failing test file simultaneously without shared state interference.

Yes, you can use parallel agents to fix separate subsystems, provided the failures are truly independent and can be examined without shared state. Each agent needs clear constraints to prevent unintended changes or interference during the concurrent investigation.

To prevent interference when dispatching multiple agents, provide each agent with a self-contained prompt, narrow scope, and clear constraints. After all agents complete, collect their summaries, check for conflicts, and run the full test suite before integrating fixes.

You should not use parallel agents for debugging when test failures share state or depend on each other, as the Skill requires independent problem domains to run concurrent investigations without interference or conflict during integration.

Yes, dispatching parallel agents requires post-run review with full test-suite verification before integration. After collecting concise summaries from each agent and checking for conflicts, you run the complete CI pipeline to validate the non-conflicting fixes.