Monitors deployed URLs after releases for regressions such as HTTP errors, console errors and performance drops.
- What it does
- Canary Watch checks a deployed URL for regressions after a deployment, merge or dependency upgrade. It inspects HTTP status, console errors, failed network calls, performance metrics (LCP/CLS/INP), key content elements and API health against a baseline. It supports a single quick check, continuous monitoring at an interval for a duration, and a diff mode comparing staging with production. It reports results in a markdown canary report and can raise desktop or webhook alerts when critical thresholds are exceeded.
- When to use it
- Use it after deploying to production or staging, after merging a high-risk PR, when verifying that a fix works, during a release window, or after a dependency upgrade.
- Requirements
- Instructions only; no scripts are shipped. It relies on the agent's browser and network access to reach the monitored URLs, and optionally desktop notifications or Slack/Discord webhooks for alerts. It writes a log to ~/.claude/canary-watch.log.