What problem does it solve? Deploys that pass CI can still break in production due to missing environment variables, stale CDN caches, or slow migrations. This Skill watches the live app in the first minutes after a deploy and catches regressions before users report them. ## Core Features & Use Cases - Baseline Capture: Take pre-deploy screenshots, console error counts, and load times per page, then compare post-deploy state against them. - Continuous Monitoring Loop: Check pages every 60 seconds for a configurable duration (1-30 minutes), alerting only on changes that persist across consecutive checks. - Severity-Based Alerts: Classify findings as CRITICAL (page load failure), HIGH (new console errors), MEDIUM (2x load time regression), or LOW (new broken links), each with screenshot evidence. - Use Case: After shipping a release, run the canary against your production URL for 10 minutes. It auto-discovers navigation pages, watches them, and produces a health report with a HEALTHY, DEGRADED, or BROKEN verdict. ## Quick Start Ask the AI to run /canary against your production URL to monitor the deploy for 10 minutes and report any console errors or performance regressions.
