eval__implement-proof-gate

A proof harness validates an approved objective gate by generating and testing required artifacts before product behavior is implemented. It captures fresh artifacts, updates the eval plan, and yields READY_RED or READY_GREEN classifications with a clear artifact trail.

You can validate an eval gate without modifying product code by running a generated harness that exercises the gate's pass/fail criteria independently. This approach strictly proves the gate and syncs results to eval-plan.md.

Gate validation in software engineering ensures an objective meets concrete pass/fail criteria. You need a proof harness when product behavior is not yet implemented, allowing you to capture fresh artifacts and establish a clear pass or fail classification.

No, a proof harness applies specifically to gates with concrete pass/fail criteria. The harness requires distinct pass or fail conditions to generate test artifacts and accurately classify the gate as READY_GREEN or READY_RED.

The best way to sync test artifacts to an eval plan is by running a proof harness that automatically captures fresh artifacts and updates eval-plan.md. This ensures the plan reflects the latest harness details and gate classifications.

A proof harness returns a READY_RED classification when the generated test artifacts fail to meet the approved objective gate's pass/fail criteria. This indicates the gate is not satisfied and requires product behavior adjustments before re-evaluation.