orchestration-guard — Mission Control

← Back to Tasks

OVERRIDES runtime profile: uses direct_python (plain Python, no model) because the script only reads mc.db read-only, runs git evidence checks, and sends a Telegram alert via notify.py — it never invokes the claude CLI or any LLM API.

Manual run / dry-run:

python3 ~/workspace/scripts/orchestration_guard.py --hours 24 --dry-run

Schedule Every 6 hours at :3030 */6 * * *

EnabledYes

Run Asscript

Timeout600s

RetryNo

Next Runin 19m (2026-06-13 06:30)

Task Definition

tasks/orchestration-guard.md

Run Asscript

Commandset

Tool-lessNo

BackgroundNo

Runtime profile

Notify destination

python3 ~/workspace/scripts/orchestration_guard.py --hours 12

---
id: orchestration-guard
name: Orchestration guard — false-done + Needs-you purity monitor
schedule: "30 */6 * * *"
enabled: true
criticality: medium
run_as: script
command: python3 ~/workspace/scripts/orchestration_guard.py --hours 12
description: Standing monitor from the 2026-06-10 orchestration reliability audit (items B+F). Two deterministic checks against mc.db, no LLM. (F) False-done lifecycle — tickets closed 'done' in the last 12h that had a real worker (task_run/worker_pid/session) on a code project but show NO landed-commit evidence (done_sha unreachable from master AND no master commit mentioning the ticket id — cherry-pick safe). (B) Needs-you purity — dev/MC tickets parked on Elmar containing coding questions that the orchestrator should have decided. Writes ~/workspace/reports/orchestration-guard/YYYY-MM-DD.md; Telegram-alerts ONLY on violations (clean runs silent). Complements the preventative MC-4913 close gate as the detective control.
runtime_profile: direct_python
notify_to: elmar
---

**OVERRIDES runtime profile:** uses `direct_python` (plain Python, no model) because the script only reads mc.db read-only, runs git evidence checks, and sends a Telegram alert via notify.py — it never invokes the `claude` CLI or any LLM API.

Manual run / dry-run:

```bash
python3 ~/workspace/scripts/orchestration_guard.py --hours 24 --dry-run
```

Edit Schedule

Frequency

Result:

Runs (7d)

Completed

Failed

Timeout

100%

Success (7d)

Avg Duration

Run History (12)

Started	Finished	Status	Duration	Output	Log
6h ago	6h ago	completed	1s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-13.md (1 violation(s))	View log
12h ago	12h ago	completed	1s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-12.md (4 violation(s))	View log
18h ago	18h ago	completed	1s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-12.md (3 violation(s))	View log
24h ago	24h ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-12.md (2 violation(s))	View log
1d ago	1d ago	completed	1s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-12.md (2 violation(s))	View log
1d ago	1d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-11.md (0 violation(s))	View log
2d ago	2d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-11.md (0 violation(s))	View log
2d ago	2d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-11.md (0 violation(s))	View log
2d ago	2d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-11.md (1 violation(s))	View log
2d ago	2d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-10.md (5 violation(s))	View log
3d ago	3d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-10.md (8 violation(s))	View log
3d ago	3d ago	completed	0s	report: /home/lucienne/workspace/reports/orchestration-guard/2026-06-10.md (0 violation(s))	View log

Comments (0)

No comments yet — add instructions for Luci

←