[detection-stats] Detection stats for github/gh-aw - 2026-09-13

601 external-detector runs analysed, detection-job error rate 1%, soft failures 0, threat rate 0% (of 145 runs with a verdict).

## Summary

# gh-aw detection statistics - 2026-09-13 (UTC)

Repository: `github/gh-aw`
Window: `2026-09-13T00:00:00Z` .. `2026-09-13T23:59:59Z`
API requests: 1512, rate-limit pauses: 1
Data complete: yes

## Totals

| Metric | Count |
|---|---|
| Workflow runs in window | 1423 |
| Agentic runs (`*.lock.yml`) | 872 |
| Runs with a `detection` job | 678 |
| ... using the external detector | 601 |
| ... using the built-in detector | 77 |
| ... detector could not be determined | 0 |
| Agentic runs without a `detection` job | 194 |

All rates below are over the **external detector** population (601 runs). Since gh-aw #54111 the external detector is the compile-time default; a run counts as external when its `detection` job showed the `Install threat-detect binary` step, when another run of the same workflow did that day, or when the workflow is `.lock.yml` and no run showed the built-in shape (a completed detection job with steps but no marker). Runs on a workflow that opted out with `features: gh-aw-detection: false` count as built-in.

## Detection job outcomes

| Outcome | Count | Rate |
|---|---|---|
| `success` | 305 | 50.75% |
| `skipped` | 290 | 48.25% |
| `failure` | 6 | 1% |

**Error rate (failure/timed_out/action_required): 1%**

## Verdict availability

| State | Meaning | Count |
|---|---|---|
| `present` | detection artifact downloaded and parsed | 145 |
| `skipped` | detection job was skipped or was still running (nothing to fetch) | 290 |
| `unreadable` | artifact zip could not be unpacked | 166 |

Green detection jobs that published no verdict: **0** (detection steps are `continue-on-error`, so a missing verdict artifact is the only reliable signal for these).

## Detection results

| Result | Count |
|---|---|
| Runs with a parsed verdict | 145 |
| Clean (no threat) | 145 |
| Any threat | 0 |
| `prompt_injection` | 0 |
| `secret_leak` | 0 |
| `malicious_patch` | 0 |

**Threat rate (of runs with a verdict): 0%**

## By workflow

| Workflow | Runs | Failed | Cancelled | Skipped | No verdict | Threats |
|---|---|---|---|---|---|---|
| PR Sous Chef | 80 | 0 | 0 | 51 | 0 | 0 |
| Daily Trajectory Grader Implementer | 46 | 0 | 0 | 45 | 0 | 0 |
| Issue Monster | 46 | 0 | 0 | 2 | 0 | 0 |
| [aw] Failure Investigator (6h) | 46 | 0 | 0 | 43 | 0 | 0 |
| Deployment Incident Monitor | 44 | 0 | 0 | 44 | 0 | 0 |
| Avenger | 24 | 0 | 0 | 15 | 0 | 0 |
| Q | 23 | 0 | 0 | 23 | 0 | 0 |
| Test Quality Sentinel | 14 | 0 | 0 | 0 | 0 | 0 |
| Daily Go Test Parallelizer | 12 | 0 | 0 | 1 | 0 | 0 |
| Design Decision Gate 🏗️ | 12 | 0 | 0 | 0 | 0 | 0 |
| Impeccable Skills Reviewer | 12 | 0 | 0 | 0 | 0 | 0 |
| Matt Pocock Skills Reviewer | 12 | 0 | 0 | 0 | 0 | 0 |
| PR Code Quality Reviewer | 12 | 0 | 0 | 0 | 0 | 0 |
| Ponytail Reviewer | 12 | 0 | 0 | 0 | 0 | 0 |
| Auto-Triage Issues | 7 | 0 | 0 | 0 | 0 | 0 |
| Contribution Check | 5 | 0 | 0 | 0 | 0 | 0 |
| Code Scanning Fixer | 4 | 0 | 0 | 0 | 0 | 0 |
| Deep Report | 4 | 0 | 0 | 0 | 0 | 0 |
| PR Triage Agent | 4 | 0 | 0 | 0 | 0 | 0 |
| Squad — ``@copilot`` run pr-finisher skill | 4 | 0 | 0 | 4 | 0 | 0 |
| Smoke Aider | 3 | 1 | 0 | 2 | 0 | 0 |
| Smoke Crush | 3 | 1 | 0 | 2 | 0 | 0 |
| Smoke DeepSeek Harness | 3 | 1 | 0 | 2 | 0 | 0 |
| Smoke Gemini | 3 | 0 | 0 | 2 | 0 | 0 |
| Smoke Goose | 3 | 1 | 0 | 2 | 0 | 0 |

_151 further workflows omitted; see `stats.json`._

## Notable runs

_(106 further notable runs omitted; see `stats.json` on the collection run.)_

## Change since 2026-09-12

| Metric | 2026-09-12 | 2026-09-13 | Delta | 7-day mean |
|---|---|---|---|---|
| External detector runs | 598 | 601 | +3 | 905.86 |
| Error rate | 0.0% | 1% | +1.0pp | 0.06% |
| Soft failures | 11 | 0 | -11 | 80.14 |
| Runs with verdict | 95 | 145 | +50 | 234.0 |
| Any threat | 0 | 0 | 0 | 0.14 |
| Threat rate | 0.0% | 0% | 0pp | 0.05% |

## Watch list

- Smoke Aider — 1 failed, 0 without a verdict, out of 3 runs
- Smoke Crush — 1 failed, 0 without a verdict, out of 3 runs
- Smoke DeepSeek Harness — 1 failed, 0 without a verdict, out of 3 runs
- Smoke Goose — 1 failed, 0 without a verdict, out of 3 runs
- Smoke OpenCode — 1 failed, 0 without a verdict, out of 3 runs

Collected by: https://github.com/github/gh-aw-threat-detection/actions/runs/34804263490
Full data: the detection-stats-34804263490 artifact on that run.

> Generated by [Detection Stats Daily](https://github.com/github/gh-aw-threat-detection/actions/runs/34804263490) · copilot · auto · 24.3 AIC · ⌖ 4.12 AIC · ⊞ 10K · [◷](https://github.com/search?q=repo%3Agithub%2Fgh-aw-threat-detection+is%3Aissue+%22gh-aw-workflow-call-id%3A+github%2Fgh-aw-threat-detection%2Fdetection-stats-daily%22&type=issues)

This issue is an automatically generated daily detection report rather than a requested change. Start with the linked Detection Stats Daily run and its detection-stats-34804263490 artifact, including stats.json; the payload gives no file, test, or completion criteria for follow-up work.

Written by the indexing model from the issue text.