mphaxise Canary
Use this skill after deployment when the right move is to watch the live app for regressions instead of assuming success.
This port is adapted from garrytan/gstack at commit 4d2c8d94d00cc4f4f3d4c26316a4f939ceedc045.
Workflow
- Define the live surface to monitor.
- Check the most important production signals:
- page availability
- console or runtime errors
- key route health
- performance drift
- Record the result in a report if the run should be compared later.
- If recurring checks are useful, hand off to
cron-scheduler.
Guardrails
- Do not treat one successful page load as comprehensive health.
- Be explicit about what was not monitored.
- Prefer short, repeatable checks over broad but vague post-deploy confidence.