Last released Jul 25, 2026
A code review gate that verifies changes by executing tests, not by judging diffs. An LLM proposes edge-case tests; a deterministic harness runs each one; a behavior is a gap only if its test runs and fails.
Supported by
