reduce-orchestrator

You can orchestrate parallel map-worker tasks that concurrently analyze different log files, then synthesize their narrative reports into a decision to iterate or finish, ensuring robust debugging outcomes.

A MapReduce workflow with verification coordinates parallel map-worker tasks, synthesizes narrative-first reductions, and enforces a mandatory parallel Verify phase to check critical findings before deciding on a fix.

Deterministic state management persists all run state in a `.rlm/` directory, enforcing strict contract compliance and concurrency safety to ensure reproducible and robust research outcomes.

Yes, the orchestrator coordinates parallel map-worker tasks under optional hint_paths, guiding workers while synthesizing their narrative reports into a final decision to iterate or finish.

Automating multi-step research tasks requires orchestrating parallel workers, synthesizing their results, and performing a mandatory verification phase to ensure reproducible and robust outcomes.

You should avoid MapReduce orchestration for simple, single-threaded analysis tasks that do not require coordinating numerous concurrent workers, synthesizing narrative reports, or a dedicated verification phase.