dispatching-parallel-agents

Parallel investigation dispatches separate agents to debug independent failures concurrently across different files or subsystems. It assigns one agent per problem domain to work simultaneously, reducing total turnaround time without requiring shared state between them.

Dispatch parallel agents when you have 3 or more independent failures across files or subsystems where each domain can be investigated without context from others. It requires no shared state between agents and outputs a consolidated summary for integration.

The best way to coordinate concurrent debugging is independent-domain dispatch, assigning one agent per problem domain to work concurrently. It enables simultaneous exploration of unrelated failures and returns summarized findings to be reviewed and merged safely.

Parallel agent orchestration cannot be used if your debugging tasks require shared state between subsystems. It strictly applies to 3 or more independent failures where each domain can be investigated without context from others.

To start dispatching parallel agents for independent task investigation, provide 3 independent tasks to be worked on in parallel and collect their summaries. The agents will return summarized findings to be reviewed and merged safely.