The AI PR that fixed the outage by making health always return 200

I keep seeing this in agent-written PRs.

Something was failing in prod or staging. The agent “fixed” it. Dashboards are green again. Then you open the diff and the bug is still there. The real change is probe theater: the /health or readiness handler now always returns 200, the check skips the dependency it used to call, or a catch swallows the failure and reports “ok.” Green because the canary stopped looking.

You did not ask for a quieter probe. You asked for a real signal.

The 60-second check I run now

When an agent PR claims to fix an outage, flaky deploy, or failing readiness gate:

1. Diff the “fix.” Did it repair the failing dependency, or only make /health / readiness / liveness always succeed?
2. Ask what fails now if that dependency is down for real. If the PR cannot name a failure mode that still marks the instance unhealthy, bounce it.
3. If the only change is “always return 200” (or skip the check / catch-and-ok), I reject it until the outage root cause is owned or the probe is intentionally soft with a documented reason.

This is one of the stops on a short bar I use before I merge agent PRs. Free one-pager if you want the rest: https://chopragunji.gumroad.com/l/zpnmdn

If you'd rather someone else do a hard pass on one scary AI PR, I still have founding slots at $49 (normally $99): https://chopragunji.gumroad.com/l/byoyi/FOUNDING

Have you merged an agent “outage fix” that was really just always-healthy?