[Fleet Monitor] Dev-lead stub drift #2284

Activity

1. github-actions

   added

   health-checkAutomated health check reportAutomated health check report

   dev-leadFor dev-lead agent pickupFor dev-lead agent pickup

   fleet-stub-driftPer-repo workflow stub drifted from the canonical templatePer-repo workflow stub drifted from the canonical template

   on Oct 11, 2026
2. don-petry

   added

   dev-lead:hands-offExclude from dev-lead agent automationExclude from dev-lead agent automation

   on Oct 11, 2026
3. don-petry commented on Oct 11, 2026

   

   don-petry

   on Oct 11, 2026

   Collaborator

   More actions

   dev-lead is withholding action on this item.

   It is labeled dev-lead:hands-off (deliberately kept off-limits to dev-lead (e.g. changes to the dev-lead automation itself)), so dev-lead will not pick it up while that label is present. This notice is posted once so the withhold is visible rather than looking like a stalled run.

   To re-enable automated pickup: remove the dev-lead:hands-off label.
4. don-petry commented on Oct 12, 2026

   

   don-petry

   on Oct 12, 2026 – with Claude

   Collaborator

   More actions

   Closing this issue. It no longer describes live drift.

   • The markets row is fixed. The canonical template standards/workflows/agent-ingress.yml is now on petry-projects/.github main (452ea69, from .github#1316). markets main's ingress (9310b6e, from fix(agent-ingress): pin ci-failure-analyst to v1-stable; fix the analyst loop guard markets#523) matches it in all five job blocks: dev-lead 4f0b41d, pr-auto-review 69e0115, pr-review e14fb04, pr-review-mention d21265d and ci-failure-analyst 4368213.
   • The other rows are legacy repos. Those repos have no agent-ingress.yml. Since feat: implement issue #2289 — Fleet Monitor reports legacy per-role stub drift as agent-ingress.yml drift (alerts and remediation plan target repos that never collapsed) #2292, a repo classified from its per-role stub is shown in the report but no longer feeds fleet_stub_drift.json or this alert.
   • Evidence. In Fleet Monitor run 38193525415 (dry-run, 2026-10-12), the monitor read the five canonical block SHAs above from .github main. Its drift alert covered only the Initiative-planner and Initiative-driver stubs and updated none of the role issues. Its dry-run remediation for markets touched only those two initiative stubs and left the ingress alone.
   • ALIGNED counts. That run's report was cut off before the role sections. I re-ran scripts/fleet_monitor.sh over the same two merged files with the gh shim from tests/fleet_stub_registry.bats. It reported ALIGNED: 1 DRIFTED: 0 MISSING: 0 for all five roles and an empty fleet_stub_drift.json.

   If real ingress drift comes back, the monitor will open a new issue.

   Refs #1729

   ---

   Generated by Claude Code
5. don-petry

   closed this as completedon Oct 12, 2026