test: smoke-test the central PR-e2e pipeline (do not merge)

[bnsoni]

Throwaway PR to exercise the new central PR-e2e pipeline end to end. Do not merge — delete the branch when done.

How to run it

Add the run-tests label. That is the only trigger.

What should happen

1. gate passes (label present, not a fork)
2. unit-tests + build-app-image run — the mentor SPA image, --no-cache, from this PR's merge commit
3. The e2e job dispatches iblai/iblai-deploy-ops and streams phase progress into this log:

   [hh:mm:ssZ] resolve: success | baseline drift check: in_progress
   [hh:mm:ssZ] … | sync env → 1.116.0: in_progress      ← only if stg1 has drifted
   [hh:mm:ssZ] … | swap mentor → PR image: success | e2e: in_progress
   

4. A comment appears on this PR with the verdict and a public link to the Playwright report + traces
5. stg1's mentor SPA is restored to its original image (this always runs, even on failure)

Why this spec

The change is a comment only in one spec file, so the changed-spec detector picks selective mode and runs just 33-ui-render-console-errors.spec.ts rather than the ~50 min mentor suite. That spec is a single test that passed clean on the last full run, so a red result means the pipeline, not the app.

Expect roughly 10–15 min — most of it the Playwright image build. Longer on the first run if stg1 has to be synced to the prod release first.

What to check

• the phase log is readable here, without opening the private repo
• the PR comment lands, with a working report link
• PR E2E Tests reports a conclusion
• stg1 mentor is back on its original image afterwards

🤖 Generated with Claude Code

[mamigot]

Claude encountered an error —— View job

---

I'll analyze this and get back to you.

[github-actions]

PR E2E - OK

Full report, traces and screenshots

result	OK
suite	mentor - mode selective - env stg1
tested against	prod release 1.116.0 (env already on it)
failed	0
new failures	0