gsd-maritime-execute

Maritime GSD execute-phase coding parallelizes implementation and instant critique by triggering a coder agent to write and test code while a critic agent concurrently performs quality and safety reviews, reporting pass or fail results instantly.

The concurrent coder and critic workflow handles code safety checks by running a critic agent in parallel with implementation to perform quality and safety reviews, triggering a repair attempt for severe issues before escalating to the user.

Automating unit testing and self-checking in maritime robotics development is best achieved by using a coding agent that implements task requirements, runs unit tests, and performs self-checks before a parallel critic agent reviews the output.

Yes, you need to provide a task specification to start the parallel review and coding process, which triggers the maritime coder and critic agents to implement, test, and review the code simultaneously.

When the critic agent rejects code during a parallel review, the main agent triggers a repair attempt to fix the severe problems, and if the repair fails, it escalates the issue to the user for manual intervention.