[detection-stats] Detection stats for github/gh-aw - 2026-09-10

811 external-detector runs analysed, detection-job error rate 0%, 84 soft failures, threat rate 0%.

## Summary

# gh-aw detection statistics - 2026-09-10 (UTC)

Repository: `github/gh-aw`
Window: `2026-09-10T00:00:00Z` .. `2026-09-10T23:59:59Z`
API requests: 1807, rate-limit pauses: 1
Data complete: yes

## Totals

| Metric | Count |
|---|---|
| Workflow runs in window | 1707 |
| Agentic runs (`*.lock.yml`) | 1052 |
| Runs with a `detection` job | 812 |
| ... using the external detector | 811 |
| ... using the built-in detector | 1 |
| ... detector could not be determined | 0 |
| Agentic runs without a `detection` job | 240 |

All rates below are over the **external detector** population (811 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` | 409 | 50.43% |
| `skipped` | 402 | 49.57% |

**Error rate (failure/timed_out/action_required): 0%**

## Verdict availability

| State | Meaning | Count |
|---|---|---|
| `present` | detection artifact downloaded and parsed | 281 |
| `absent` | detection job ran but published no artifact (soft failure) | 84 |
| `skipped` | detection job was skipped or was still running (nothing to fetch) | 402 |
| `unreadable` | artifact zip could not be unpacked | 44 |

Green detection jobs that published no verdict: **84** (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 | 281 |
| Clean (no threat) | 281 |
| 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 | 82 | 0 | 0 | 0 | 3 | 0 |
| Q | 66 | 0 | 0 | 66 | 0 | 0 |
| Daily Trajectory Grader Implementer | 47 | 0 | 0 | 47 | 0 | 0 |
| Issue Monster | 47 | 0 | 0 | 0 | 0 | 0 |
| [aw] Failure Investigator (6h) | 47 | 0 | 0 | 43 | 0 | 0 |
| Deployment Incident Monitor | 40 | 0 | 0 | 40 | 0 | 0 |
| Auto-Triage Issues | 31 | 0 | 0 | 21 | 10 | 0 |
| Avenger | 24 | 0 | 0 | 15 | 8 | 0 |
| Workflow Generator | 19 | 0 | 0 | 19 | 0 | 0 |
| Test Quality Sentinel | 15 | 0 | 0 | 0 | 2 | 0 |
| Design Decision Gate 🏗️ | 14 | 0 | 0 | 0 | 0 | 0 |
| Impeccable Skills Reviewer | 14 | 0 | 0 | 0 | 2 | 0 |
| Matt Pocock Skills Reviewer | 14 | 0 | 0 | 0 | 2 | 0 |
| PR Code Quality Reviewer | 14 | 0 | 0 | 0 | 6 | 0 |
| Ponytail Reviewer | 14 | 0 | 0 | 1 | 1 | 0 |
| Daily Go Test Parallelizer | 12 | 0 | 0 | 0 | 5 | 0 |
| Contribution Check | 6 | 0 | 0 | 0 | 0 | 0 |
| Code Scanning Fixer | 4 | 0 | 0 | 0 | 1 | 0 |
| Deep Report | 4 | 0 | 0 | 0 | 0 | 0 |
| PR Triage Agent | 4 | 0 | 0 | 0 | 1 | 0 |
| Squad — Static GitHub agent enclaves omit the safeoutputs write-sink policy | 3 | 0 | 0 | 3 | 0 | 0 |
| The Great Escapi | 3 | 0 | 0 | 2 | 0 | 0 |
| Agent Container Smoke Test | 2 | 0 | 0 | 2 | 0 | 0 |
| Changeset Generator | 2 | 0 | 0 | 2 | 0 | 0 |
| Code Refiner | 2 | 0 | 0 | 2 | 0 | 0 |

_248 further workflows omitted; see `stats.json`._

## Notable runs

_See the `detection-stats-` artifact for the full notable-runs list (68 further entries omitted here); all sampled notable runs on 2026-09-10 had `success` job conclusions with `absent` or `unreadable` verdicts and no reported threats._

## Change since 2026-09-09

| Metric | 2026-09-09 | 2026-09-10 | Delta | 7-day mean |
|---|---|---|---|---|
| External detector runs | 853 | 811 | -42 | 829.57 |
| Error rate | 0% | 0% | 0 pp | 0% |
| Soft failures | 79 | 84 | +5 | 87.43 |
| Runs with verdict | 318 | 281 | -37 | 267.57 |
| Any threat | 0 | 0 | 0 | 0.14 |
| Threat rate | 0% | 0% | 0 pp | 0.05% |

## Watch list

- Auto-Triage Issues — 0 failed, 10 without a verdict, out of 31 runs
- Avenger — 0 failed, 8 without a verdict, out of 24 runs
- PR Code Quality Reviewer — 0 failed, 6 without a verdict, out of 14 runs
- Daily Go Test Parallelizer — 0 failed, 5 without a verdict, out of 12 runs
- PR Sous Chef — 0 failed, 3 without a verdict, out of 82 runs

Collected by: https://github.com/github/gh-aw-threat-detection/actions/runs/34560277413
Full data: the detection-stats-34560277413 artifact on that run.

> Generated by [Detection Stats Daily](https://github.com/github/gh-aw-threat-detection/actions/runs/34560277413) · copilot · auto · 22.4 AIC · ⌖ 5.56 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)

Start with the linked Detection Stats Daily workflow run and its detection-stats-34560277413 artifact; the issue identifies stats.json as the full data source. There is no requested change or stated acceptance condition, so completion cannot be determined from this issue alone.

Written by the indexing model from the issue text.