Repair failing pull-request CI gate so director merges can read it green #424

New issue

Closed

opened 2026-07-10 09:39:57 +00:00 by coilyco-ops · 4 comments

coilyco-ops commented

2026-07-10 09:39:57 +00:00

Member

Problem

agentic-os#423 adds a ci / gate (pull_request) workflow for agentic-os#421, but the live status for head a693bbc0df9e63d6a73951279dce2b9310666d74 is already failure:

context: ci / gate (pull_request)
status: failure
description: Failing after 7s
target: /coilyco-flight-deck/agentic-os/actions/runs/885/jobs/0

The director sees the context but still skips the PR:

skipping coilyco-flight-deck/agentic-os#423: linked PR head SHA a693bbc0df9e63d6a73951279dce2b9310666d74 has required status context ci / gate (pull_request)=

A review comment was posted on PR #423, but the running agentic-os#421 engineer will not re-read comments. This fresh issue is the repair channel.

Acceptance criteria

The new pull-request workflow runs green on its own PR.
The workflow uses the repo's ward verbs or a verified fallback that exists in the selected container image.
The director dry-run can read the context as passing once the status is green.
The repair preserves a pull_request context named for branch protection, currently ci / gate (pull_request).
PR #423 is updated or replaced so agentic-os#421 can close only when the gate works end to end.

## Problem `agentic-os#423` adds a `ci / gate (pull_request)` workflow for `agentic-os#421`, but the live status for head `a693bbc0df9e63d6a73951279dce2b9310666d74` is already `failure`: ```text context: ci / gate (pull_request) status: failure description: Failing after 7s target: /coilyco-flight-deck/agentic-os/actions/runs/885/jobs/0 ``` The director sees the context but still skips the PR: ```text skipping coilyco-flight-deck/agentic-os#423: linked PR head SHA a693bbc0df9e63d6a73951279dce2b9310666d74 has required status context ci / gate (pull_request)= ``` A review comment was posted on PR #423, but the running `agentic-os#421` engineer will not re-read comments. This fresh issue is the repair channel. ## Acceptance criteria * The new pull-request workflow runs green on its own PR. * The workflow uses the repo's ward verbs or a verified fallback that exists in the selected container image. * The director dry-run can read the context as passing once the status is green. * The repair preserves a `pull_request` context named for branch protection, currently `ci / gate (pull_request)`. * PR #423 is updated or replaced so `agentic-os#421` can close only when the gate works end to end.

coilyco-ops added the

headless

labels

2026-07-10 09:39:57 +00:00

coilyco-ops commented

2026-07-10 09:40:14 +00:00

Author

Member

WARD-DISPATCH: deferred ⏸

deferred details

This forwarded dispatch was deferred after the issue was already reserved.

Attempted harness: codex
Attempted run: ward agent engineer coilyco-flight-deck/agentic-os#424 --harness codex --details Fix agentic-os#424. PR #423 creates ci / gate but the live pull_request status fails after 7s and director dry-run still skips it. Diagnose the workflow/container failure, update or replace PR #423, prove the PR gate runs green and remains director-readable, then close agentic-os#424.
Container: engineer-codex-agentic-os-424
Container created: no running engineer was observed.
Host log: /Users/kai/.ward/agent-logs/dispatch/20260710T094012Z-director-codex-zv99-coilyco-flight-deck-agentic-os-424.log
Capacity: ward agent engineer --harness codex: global engineer limit is reached: 12 running (limit 12); wait for a run to finish or run ward agent reap for stale engineers

Retry: the issue stays queued and the director will try again when a slot opens.

— Codex, via ward agent

WARD-DISPATCH: deferred ⏸ <details><summary>deferred details</summary> This forwarded dispatch was deferred after the issue was already reserved. Attempted harness: `codex` Attempted run: `ward agent engineer coilyco-flight-deck/agentic-os#424 --harness codex --details Fix agentic-os#424. PR #423 creates ci / gate but the live pull_request status fails after 7s and director dry-run still skips it. Diagnose the workflow/container failure, update or replace PR #423, prove the PR gate runs green and remains director-readable, then close agentic-os#424.` Container: `engineer-codex-agentic-os-424` Container created: no running engineer was observed. Host log: `/Users/kai/.ward/agent-logs/dispatch/20260710T094012Z-director-codex-zv99-coilyco-flight-deck-agentic-os-424.log` Capacity: `ward agent engineer --harness codex: global engineer limit is reached: 12 running (limit 12); wait for a run to finish or run `ward agent reap` for stale engineers` Retry: the issue stays queued and the director will try again when a slot opens. </details>  — Codex, via `ward agent`

coilyco-ops commented

2026-07-10 09:44:01 +00:00

Author

Member

Additional evidence for the repair:

PR #423 advanced to head 29d53be09e30298aac1a87ceb9f7d2d1b0653885.
Forgejo status history for that head still reports ci / gate (pull_request) as failure, description Failing after 6s, target /coilyco-flight-deck/agentic-os/actions/runs/886/jobs/0.
Director dry-run still reports skipping coilyco-flight-deck/agentic-os#423: linked PR head SHA 29d53be09e30298aac1a87ceb9f7d2d1b0653885 has required status context ci / gate (pull_request)=.
The current .ward/forgejo-actions-logs.sh route returns 404 for run 886 job 0, so the repair likely needs to diagnose via the Forgejo UI or fix the Actions log surface too.

Additional evidence for the repair: * PR `#423` advanced to head `29d53be09e30298aac1a87ceb9f7d2d1b0653885`. * Forgejo status history for that head still reports `ci / gate (pull_request)` as `failure`, description `Failing after 6s`, target `/coilyco-flight-deck/agentic-os/actions/runs/886/jobs/0`. * Director dry-run still reports `skipping coilyco-flight-deck/agentic-os#423: linked PR head SHA 29d53be09e30298aac1a87ceb9f7d2d1b0653885 has required status context ci / gate (pull_request)=`. * The current `.ward/forgejo-actions-logs.sh` route returns 404 for run `886` job `0`, so the repair likely needs to diagnose via the Forgejo UI or fix the Actions log surface too.

coilyco-ops commented

2026-07-10 09:53:00 +00:00

Author

Member

Additional current-head evidence:

PR #423 head 4a689cae78a20ed3cba7a123cf8bc5ed650b2481 now has ci / gate (pull_request) = failure, description Failing after 5m15s, target /coilyco-flight-deck/agentic-os/actions/runs/888/jobs/0.
Earlier heads a693bbc... and 29d53be... also failed. The repair should diagnose the actual job failure, not just rotate between equivalent workflow forms.

Additional current-head evidence: * PR `#423` head `4a689cae78a20ed3cba7a123cf8bc5ed650b2481` now has `ci / gate (pull_request)` = `failure`, description `Failing after 5m15s`, target `/coilyco-flight-deck/agentic-os/actions/runs/888/jobs/0`. * Earlier heads `a693bbc...` and `29d53be...` also failed. The repair should diagnose the actual job failure, not just rotate between equivalent workflow forms.

coilyco-ops commented

2026-07-10 10:17:14 +00:00

Author

Member

2026-07-10 follow-up: PR #423 has now merged, and its final head 3d50ba28b4f83661775320f7e9f13d644c620f70 reports ci / gate (pull_request) with status=success, description Successful in 11s, target /coilyco-flight-deck/agentic-os/actions/runs/889/jobs/0. Closing this repair issue as fixed by #423. The remaining director status-read bug is tracked separately in coilyco-flight-deck/ward#994.

2026-07-10 follow-up: PR #423 has now merged, and its final head `3d50ba28b4f83661775320f7e9f13d644c620f70` reports `ci / gate (pull_request)` with `status=success`, description `Successful in 11s`, target `/coilyco-flight-deck/agentic-os/actions/runs/889/jobs/0`. Closing this repair issue as fixed by #423. The remaining director status-read bug is tracked separately in coilyco-flight-deck/ward#994.

coilyco-ops closed this issue

2026-07-10 10:17:14 +00:00