adversarial-verify

Automate QA verification by re-running findings and checks from a QA manifest, performing pixel-diff comparisons against baselines, and generating a structured verification-report.yaml alongside cycle evidence.

Adversarial QA verification re-evaluates findings across test cycles to detect regressions. You need it when re-running manifests to produce pixel-diff evidence, API responses, and test outputs for comprehensive regression checks.

Yes, you need Playwright and pixelmatch tooling, along with a running backend and frontend, to execute the QA manifest re-runs and produce the pixel-diff images required for the verification-report.yaml.

Run the verification command with your manifest slug to re-run findings and checks. The process generates cycle evidence, including screenshots and pixel-diff images, and outputs a formal verification-report.yaml summarizing results.

The best way is to re-run checks from an adversarial QA manifest, capturing screenshots and API responses as cycle evidence. This performs pixel-diff comparisons against baselines to detect regressions and outputs a verification-report.yaml.

QA verification requires a running backend and frontend because the process re-runs checks from a QA manifest against live endpoints, capturing API responses and generating screenshots for pixel-diff regression checks.