What problem does it solve? It turns vague claims like "did this fix it" or "prove it works" into falsifiable statements backed by repeatable local evidence, replacing subjective recaps with measured before-and-after comparisons. ## Core Features & Use Cases - Baseline/Treatment Comparison: Captures evidence from the old state (merge base, parent commit, failing branch) and the changed state using identical commands, data, and environment. - Multiple Verification Surfaces: Supports code tests, CLI transcripts, UI screenshots, API response diffs, performance timings, and memory heap snapshots. - Strict Verdicts: Returns exactly one of VERIFIED, NOT VERIFIED, or INCONCLUSIVE with raw metrics and reasoning. - Use Case: After fixing a performance regression, ask the assistant to verify the fix; it measures the parent commit and the patched code with the same benchmark and reports whether the claimed threshold was met. ## Quick Start Ask the assistant to verify that your recent fix actually resolved the bug by comparing behavior before and after the change.
