bernstein's orchestrator is pure python: tick loop, task store, scheduler, router, merge queue, quality gates. no llm call sits on the coordination path. given the same backlog, the same .sdd/ state, and the same model outputs, a re-run produces the same task ordering and the same merge sequence. that determinism is what makes the hmac audit chain meaningful (no model nondeterminism in the metadata) and what lets the eval harness replay an entire run from a recorded fixture. bernstein verify --determinism <run-id> prints the execution fingerprint for a run. frameworks that put an llm in the coordination path trade this property for flexibility at planning time.
canonical answer