Detect post-deployment regressions by comparing Web Vitals and error logs against stored baselines.
npx skills add https://github.com/JYHTHEGREAT/jyh-system --skill canary-jyhthegreat
Please help me install this Agent Skill. Skill: canary Source: https://github.com/JYHTHEGREAT/jyh-system/tree/main/skills/canary Command: npx skills add https://github.com/JYHTHEGREAT/jyh-system --skill canary-jyhthegreat
After a deployment, teams often miss subtle performance degradations, console errors, or HTTP failures because they lack automated monitoring that compares current metrics against a known good baseline.
Execute /canary https://my-app.vercel.app --baseline to store a baseline, then run /canary https://my-app.vercel.app to compare and receive regression alerts.
High-intent search queries and answers about installing and using this skill.
To detect Web Vitals regressions after a Vercel deployment, capture a baseline snapshot of your production URL's metrics, console errors, and HTTP logs. Comparing subsequent live metrics against this stored baseline issues tiered alerts when regressions persist.
Yes, you can automate post-deployment monitoring using optional cron scheduling. This enables continuous polling every 60 seconds for up to 7 days, automatically re-running regression checks against your stored Web Vitals baseline for sustained alert generation.
The best way to set a baseline for production Web Vitals is executing a baseline capture command on your target Vercel production URL. This stores a reference snapshot of current metrics, console messages, and HTTP error statistics for later regression comparison.
Yes, you need browser-devtools and Vercel MCP integrations to monitor console errors and HTTP logs. The regression detection relies on these connections to collect live Web Vitals, with fallback handling available for unauthenticated Vercel or disconnected browser-devtools environments.
Tiered alerts (CRITICAL, HIGH, MEDIUM, LOW) are generated when detecting post-deployment performance regressions. These alerts trigger only when live Web Vitals, console errors, or HTTP failure metrics show persistent degradation compared to the stored baseline.
Fallback handling for unauthenticated Vercel MCP or disconnected browser-devtools guides you through alternative connection setups. It also supports optional Supabase log integration to ensure your post-deployment Web Vitals and HTTP error monitoring continues functioning.
archibateRun agent-browser and Chrome inside Vercel Sandbox microVMs for automated UI tasks.
Jia-Hong-PengAutomate end-to-end QA testing and bug remediation for web applications.
xiaohei-infoNormalize log signatures and track active issues to deduplicate alerts.
Arnutt-NTrack real-time performance metrics for HR-IMS applications with threshold alerts.
