dx-swe-parallel-agents

Group failures by independent problem domain, then dispatch one agent per domain with a focused prompt containing the failing test names, error messages, constraints, and expected output. Review each agent's summary and run the full test suite after integration.

Use parallel agents when failures are independent, such as different test files or subsystems with no shared state. Avoid parallel dispatch when failures are related, when full system context is needed, or when agents would edit the same files.

A good agent prompt is focused on one problem domain, self-contained with pasted error messages and test names, explicit about constraints like not changing production code, and specific about the expected output such as a summary of root cause and fixes.

Conflicts happen when agents share state, such as editing the same files or resources. Prevent this by scoping each agent to a distinct test file or subsystem and verifying changes do not overlap before integration.

Parallel dispatch does not fit exploratory debugging where the root cause is unknown, related failures where one fix resolves others, or situations requiring full system context. Agents can also make systematic errors, so spot-check their work.