What problem does it solve? Running regression, smoke, or sanity suites in CI is only half the job — teams still need to monitor long GitHub Actions runs, download Allure and Playwright artifacts, decide whether each failure is a real regression or a flake, and turn all of that into a defensible release decision. This Skill automates that entire Execute → Analyze → Report pipeline so release readiness is based on evidence, not guesswork. ## Core Features & Use Cases - CI Suite Orchestration: Triggers regression, smoke, or sanity workflows via gh workflow run, monitors runs to completion with background subagents, and downloads Allure, evidence, and Playwright artifacts in parallel. - Failure Classification: Applies a six-bucket decision tree (REGRESSION, FLAKY, KNOWN ISSUE, ENVIRONMENT, NEW TEST, KNOWN-BLOCKED) with a >20% flakiness rule over historical runs, excluding blocked tests from the gating pass rate. - GO / CAUTION / NO-GO Verdicts: Computes a weighted release score with hard vetoes (any @critical failure, HIGH/CRITICAL regression, or pass rate below 90%) and files confirmed defects in Jira with the mandatory field matrix. - Use Case: Before a production deploy, ask the assistant to run the staging regression suite; it triggers the workflow, waits out the 20–60 minute run, classifies every failure, and delivers an executive quality report with a release verdict saved under .context/reports/. ## Quick Start Ask the assistant to run the regression suite on staging and tell you whether the build is ready to release.
