Monitors production deployments with browser or HTTP health checks and reports verdicts.
npx skills add https://github.com/greglas75/zuvo --skill canary-greglas75
Please help me install this Agent Skill. Skill: canary Source: https://github.com/greglas75/zuvo/tree/main/skills/canary Command: npx skills add https://github.com/greglas75/zuvo --skill canary-greglas75
High-intent search queries and answers about installing and using this skill.
Run the canary skill with your production URL, for example with a 10 minute duration and 60 second interval. It performs repeated health checks for HTTP status, console errors, and response time, then reports a HEALTHY, DEGRADED, or BROKEN verdict.
When Playwright or Chrome DevTools tools are available, the skill captures Error-level console messages on every check cycle. Each new error counts as a HIGH alert and accumulates toward the --max-errors threshold that determines the final verdict.
Yes. If no browser automation tools are detected, the skill switches to degraded mode using curl for HTTP status and response time checks. Console error capture and screenshots are unavailable in this mode, and the output is annotated accordingly.
On a BROKEN verdict the skill prints a rollback command. If a deployment platform was detected it uses that platform's rollback command, otherwise it reads memory/last-ship.json for the release commit SHA and suggests git revert followed by git push.
Duration must be between 1 and 30 minutes; values outside this range stop execution with a validation error. In non-interactive environments the skill defaults to a single one-shot check unless --duration is explicitly passed.
Simon-YHKimDetects deploy platform and writes deployment configuration to CLAUDE.md for automated deploys.
svangePlan and apply repository rollbacks for faulty releases or merged PRs.
JamkrisMonitor deployed URLs for canary health across runtimes and time windows.
petrkindlmannSchedule synthetic probes across regions to monitor critical user journeys.
