MC-4411

Created from fresh MC workflow audit. This is the remaining systemic workflow issue after the UI/role-flow fixes: pickup/runtime row drift around HTTP 409 after a live runtime exists.

Luci picking up this ticket (background worker).

Ticket picked up by Luci via MC dispatcher. MC-4411: Fix ticket-pickup 409 row-drift self-healing after live runtime spawn Work this ticket in the live tmux runtime. Use DONE:, REVIEW:, or QUESTION: when you need MC to reflect the next state.

Ticket runtime send failed while dispatching. root_cause: HTTP Error 409: CONFLICT safe_retry: returned to `todo` for a fresh runtime retry. human_input_required: no

RUNTIME TERMINAL STATE (MC-3482 contract) status: warning summary: Ticket runtime send failed; parked for automatic recovery. root_cause: send failed while injecting into ticket runtime: Ticket runtime is not ready for input (status=running). Wait for the current turn to finish, or explicitly interrupt/restart the runtime before sending more terminal input. This message was not sent. safe_retry: Queued message was returned to pending and the ticket was returned to todo for a fresh runtime retry; no human input is needed. stop_condition: After the configured retry limit, leave the ticket blocked for operator inspection instead of looping. human_input_required: no next_actions: - Pickup will retry the pending message in a fresh runtime. - Inspect runtime/send only if the retry fails again. artifacts: - ticket:MC-4411

Luci picking up this ticket (background worker).

Controller normalized its own reproduction: pickup returned 409 while creating live mc-MC-4411. Ticket/runtime restored to in_progress/running with worker_pid=126269. Use this live repro as evidence for the fix.

Ticket-pickup now self-heals 409 row-drift: adopts the live runtime (in_progress + live worker_pid, PID-alive gated) instead of reverting to todo, plus a proactive reconcile pass. mc_pickup.py + test, council CLEAN, 18 tests pass, both repos pushed.

Workflow action `review_ready` applied. Work ready for review.

Gated + signed off. Verified: fix 9b5d6fd2 (workspace) + test c7d6cc1 (mission-control). _normalize_ticket_to_live_runtime adopts the live runtime on a post-spawn 409 instead of reverting to todo — breaking the todo<->409 re-claim loop. Liveness gated on the RECORDED Claude PID being alive (not just the tmux pane), so ghosts aren't resurrected; declines + lets TTL retire when PID dead. Guards confirmed: never regresses a ticket past in_progress (in_review/done), never steals from a different live worker, declines if it can't confirm ticket state. tests/test_mc4411_row_drift_selfheal.py — 8 passed locally. Council CLEAN per worker. Closing.

Shadow reviewer verdict: UNCERTAIN (advisory only). Ticket was gated to done by the orchestrator/operator — not reopening; the gate is authoritative. Reasons: The diff adds the core self-healing logic (_normalize_ticket_to_live_runtime, _reconcile_runtime_row_drift, and the 409 catch block in _send_ticket_to_tmux_runtime) with proper PID-liveness gating and guards against regressing tickets past in_progress. However, the diff is truncated mid-line in the 409 catch block, so the full integration cannot be verified. The worker claims 8 tests pass in tests/test_mc4411_row_drift_selfheal.py and a separate test commit c7d6cc1 in mission-control, but neither the test file nor the mission-control diff is included in the provided evidence, so test coverage cannot be confirmed. There is also no evidence of the required smoke/dry-run verification that ticket row, runtime_sessions, and tmux stay consistent. Gaps: - Test diff (tests/test_mc4411_row_drift_selfheal.py or mission-control commit c7d6cc1) not provided — cannot verify tests exist or cover the 409-after-spawn/live-runtime case - Diff is truncated mid-line in the 409 catch handler — full integration code not inspectable - No evidence of smoke/dry-run verification with a no-op ticket confirming consistency across ticket row, runtime_sessions, and tmux - Cannot confirm the 409 catch block actually calls _normalize_ticket_to_live_runtime or handles the adoption correctly

Luci Operator cleared stale failure/runtime state because the ticket already has review-ready worker evidence.