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
fix\(pr-pipeline\): align every needs-human path with the owner stop rule - part 3 of 6 #11021
Proposed work class: work-class: scoped
Basis: automated intake triage; the lanes App applied this class.
Size: Small
Complexity: Low
Uncertainty: Known
Flags: none
Blocked by: none
I read this as part 3 of 6 of #10517, split at intake. It asks for two changes. First, the issue lane's research-first, probe-first and split-required skips should add needs-human. Second, work-resolve-issue/report-outcome.mjs should hand patch escalations (patch-protected-path, patch-invalid and the other PATCH_REASONS) to a person. Both items are clear and can be checked by the lane. Nothing is missing. The sibling parts (#11019, #11020, #11022, #11023, #11024) cover other paths of the same parent and are not duplicates.
The work-resolve-issue lane did not start this issue: it needs a person to make it clearer first.
Why, in the run's own words:
Where should the issue lane send its three dimension skips (split-required, research-first, probe-first) instead of needs-human? The issue says to align them with the owner stop rule but does not name a lane, and the repository's own rules disagree on which lane. (a) host-run: stop-routes.json already has this route (patch-protected-path and host-only use it), and the label means 'an unattended agent run on a host'. But intake admission keeps host-run off exactly these issues on purpose: .github/scripts/intake-triage/admit.mjs:1041-1043 says host-run 'goes only on an issue whose one obstacle is that no item is left for the lane: never beside ... another refusal, a wait or a pending split', and admit.test.mjs:1632-1645 tests that Uncertainty unknown and a pending size split get no host-run. (b) Back to intake: drop status: ready and park at status: confirmed, as admit.mjs:1723-1725 does for these same refusals with no person. split-required would then go to intake's own size split (owner Q93). This needs a new stop-routes route. Also, a trusted size-split child with Size Large is admitted (admit.mjs:889-893, 'a run has no turn or time cap') but resolve-config still maps large to split (dimensions.mjs:68), so it can never leave the split-required skip. (c) fail: remove agent-ready, add no label, and the maintenance monitor reports it, which sends it to no lane. Please record which route each of the three takes. For the second done-when item, my reading of the code is that it is already aligned except patch-invalid: patch-protected-path routes to host-run, patch-not-applied and patch-apply-failed retry, and patch-empty changes nothing. patch-invalid holds because apply-issue-patch also uses it for a patch carrying a credential shape (PATCH_REASONS in report-outcome.mjs, stop-routes.json row). Can patch-invalid be recorded in the loop-lane 'When work stops for the owner' section as a kept security control, or should the write file split that reason first? That would be a change to .github/workflows/pr-run-activity-write.yml, which the issue lane cannot push.
What the run says it did not do:
Both done-when items are left until the routing question is answered; nothing was committed.
tidy: skipped. The drafted change touched the shared paths .github/ and docs/conventions/, and it was reverted.
The report-outcome and stop-routes node test suites were not run here: the lane runs no interpreter.
Whatever route is chosen, the report-skip comment in .github/workflows/work-resolve-issue.yml:410-412 ('each a hold: the human-gated label') and the dimension-skip input description in .github/workflows/pr-run-activity-write.yml:73-77 will need updating. Both are paths the issue lane may not push.
Activity
github-actions commentedon Oct 29, 2026
Proposed work class: work-class: scoped
Basis: automated intake triage; the lanes App applied this class.
Size: Small
Complexity: Low
Uncertainty: Known
Flags: none
Blocked by: none
I read this as part 3 of 6 of #10517, split at intake. It asks for two changes. First, the issue lane's
research-first,probe-firstandsplit-requiredskips should add needs-human. Second,work-resolve-issue/report-outcome.mjsshould hand patch escalations (patch-protected-path,patch-invalidand the otherPATCH_REASONS) to a person. Both items are clear and can be checked by the lane. Nothing is missing. The sibling parts (#11019, #11020, #11022, #11023, #11024) cover other paths of the same parent and are not duplicates.Automated triage, applied by the lanes App.
melodic-automation-lanes commentedon Oct 29, 2026
The work-resolve-issue lane did not start this issue: it needs a person to make it clearer first.
Why, in the run's own words:
What the run says it did not do:
Run: https://github.com/melodic-software/claude-code-plugins/actions/runs/40943532986
To retry, answer the question in a comment whose first line is
**Decision recorded.**, then removeneeds-humanand addagent-readyagain.