You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
docs(ci-policy): ci-policy.md still treats jules.yml as load-bearing after the Jules decommission #5241
Lines 160-161: describes the flow where a human applies the jules label and waits for Jules to start.
If jules.yml is gone or inert, the policy text should say so (or drop the exceptions) so future CI work doesn't reason from a dead exception. If any part of jules.yml is intentionally retained, the policy should state which part and why.
Acceptance: ci-policy.md no longer presents Jules as an active exception/flow, or explicitly documents the retained piece; no other policy terms changed.
Acceptance criteria
.github/ci-policy.md no longer presents Jules as an active exception or flow: the line-41 workflow list, the lines-152 cancelled-concurrency exception, and the lines-160-161 label-and-wait flow are removed or rewritten to state the post-decommission reality.
If any part of jules.yml is intentionally retained, ci-policy.md states which part and why; no other policy terms change.
Notes: Verified still live: no merged PR references this issue (search this run: none). One correction to the filing's premise, verified this run via the contents API: .github/workflows/jules.yml still EXISTS on main — so the fix is not simply deleting dead references; the worker must determine whether the workflow is inert (decommissioned in #5098 but file retained) or partially live, and make the policy text match that reality, per the issue's own either/or acceptance. Effort xs: a three-spot doc edit once the workflow's status is confirmed. Labels added: kind/docs, effort/xs (priority/low and status/needs-triage were already set).
— fleet monitor intake (protocol §5), 2026-11-03 20:12 CST run. Verified still live this run per the notes above.
.github/workflows/jules.yml still EXISTS but is no longer Jules: its header states it is now the pm-dispatch dispatch hub (fires a tracked OpenCode session on the laptop via the pm-dispatch starter action). The filename is intentionally retained because main-broken.yml and repo-sweeper.yml reference that path, and the jules label name is unchanged — applying it still starts remediation, now via pm-dispatch instead of the decommissioned Jules service (Jules retirement (ruled 2026-10-30) was never executed — dispatch crons, Jules-side schedules, secrets, and mcpx wiring all still live #5098).
Conclusion: the retained-piece branch of the acceptance criteria applies. The fix is a small docs-only edit to ci-policy.md (lines 41, 152, 160-161) stating the post-decommission reality: file and label retained, backing service is pm-dispatch, Jules is gone. No other policy terms change.
Activity
dezren39 commentedon Nov 4, 2026
Intake review
Related: Decommission: #5098. Policy file: .github/ci-policy.md (lines 41, 152, 160-161).
Notes: Verified still live: no merged PR references this issue (search this run: none). One correction to the filing's premise, verified this run via the contents API:
.github/workflows/jules.ymlstill EXISTS on main — so the fix is not simply deleting dead references; the worker must determine whether the workflow is inert (decommissioned in #5098 but file retained) or partially live, and make the policy text match that reality, per the issue's own either/or acceptance. Effort xs: a three-spot doc edit once the workflow's status is confirmed. Labels added: kind/docs, effort/xs (priority/low and status/needs-triage were already set).— fleet monitor intake (protocol §5), 2026-11-03 20:12 CST run. Verified still live this run per the notes above.
dezren39 commentedon Nov 4, 2026
🤖 Agent note · triaged
gh apicontents reads:.github/ci-policy.mdstill carries all three stale Jules passages the issue names: line 41 listsjules.ymlas a referenced workflow, line 152 namesjules.ymlas the cancelled-concurrency exception ("load-bearing cip=false, CI: Jules workflow_run subscribes to every workflow completion — 44 of the last 100 repo runs are skipped no-ops, and they share the concurrency group that can drop a real ci-failure-fix run #304"), lines 160-161 describe the human-applies-jules-label-and-waits-for-Jules flow..github/workflows/jules.ymlstill EXISTS but is no longer Jules: its header states it is now the pm-dispatch dispatch hub (fires a tracked OpenCode session on the laptop via the pm-dispatch starter action). The filename is intentionally retained becausemain-broken.ymlandrepo-sweeper.ymlreference that path, and thejuleslabel name is unchanged — applying it still starts remediation, now via pm-dispatch instead of the decommissioned Jules service (Jules retirement (ruled 2026-10-30) was never executed — dispatch crons, Jules-side schedules, secrets, and mcpx wiring all still live #5098).