[detection-stats] Detection stats for github/gh-aw - 2026-08-23

[github-actions]

789 external-detector runs analysed for github/gh-aw on 2026-08-23; detection-job error rate 0%; soft failures 41; threat rate 0.49% (of runs with a verdict).

Summary

gh-aw detection statistics - 2026-08-23 (UTC)

Repository: github/gh-aw
Window: 2026-08-23T00:00:00Z .. 2026-08-23T23:59:59Z
API requests: 2298, rate-limit pauses: 1
Data complete: yes

Totals

Metric	Count
Workflow runs in window	2799
Agentic runs (*.lock.yml)	1397
Runs with a detection job	800
... using the external detector	789
... using the built-in detector	11
... detector could not be determined	0
Agentic runs without a detection job	597

All rates below are over the external detector population (789 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	451	57.16%
skipped	336	42.59%
cancelled	2	0.25%

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

Verdict availability

State	Meaning	Count
present	detection artifact downloaded and parsed	408
absent	detection job ran but published no artifact (soft failure)	41
skipped	detection job was skipped or was still running (nothing to fetch)	336
unreadable	artifact zip could not be unpacked	4

Green detection jobs that published no verdict: 41 (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	408
Clean (no threat)	406
Any threat	2
prompt_injection	2
secret_leak	0
malicious_patch	0

Threat rate (of runs with a verdict): 0.49%

Reasons reported by gh-aw

From the [aw] Detection Runs tracking issue (warning/failure conclusions only), restricted to runs in the external-detector population above.

Reason	Count
agent_failure	45
threat_detected	2

By workflow

Workflow	Runs	Failed	Cancelled	Skipped	No verdict	Threats
Q	89	0	0	88	0	0
PR Sous Chef	85	0	0	8	0	0
Deployment Incident Monitor	72	0	0	72	0	0
Issue Monster	46	0	0	30	0	0
Test Quality Sentinel	33	0	1	0	1	1
Matt Pocock Skills Reviewer	32	0	0	0	1	0
Impeccable Skills Reviewer	31	0	0	0	1	0
PR Code Quality Reviewer	31	0	0	0	2	0
Ponytail Reviewer	31	0	0	0	1	0
Design Decision Gate 🏗️	30	0	0	0	4	0
Avenger	24	0	0	23	1	0
PR Description Updater	15	0	0	0	0	0
Daily Go Test Parallelizer	12	0	0	2	3	0
Auto-Triage Issues	9	0	0	1	0	0
Squad — /smoke-copilot	7	0	0	7	0	0
Contribution Check	6	0	1	0	0	0
Squad — @copilot run pr-finisher skill	5	0	0	5	0	0
Workflow Generator	5	0	0	5	0	0
Code Scanning Fixer	4	0	0	0	2	0
Deep Report	4	0	0	0	0	0
PR Triage Agent	4	0	0	0	0	0
[aw] Failure Investigator (6h)	4	0	0	0	1	0
Issue Arborist	3	0	0	0	2	0
Daily Credit Limit Test	2	0	0	1	1	0

199 further workflows omitted; see stats.json.

Change since 2026-08-22

Metric	2026-08-22	2026-08-23	Δ	7-day mean
External detector runs	962	789	-173	710.0
Error rate (%)	0.0	0.0	0.0	0.09
Soft failures	71	41	-30	42.71
Runs with verdict	429	408	-21	404.71
Any threat	2	2	0	2.57
Threat rate (%)	0.47	0.49	+0.02	0.65

Watch list

• Design Decision Gate 🏗️ — 0 failed, 4 without a verdict, out of 30 runs
• PR Code Quality Reviewer — 0 failed, 2 without a verdict, out of 31 runs
• Issue Arborist — 0 failed, 2 without a verdict, out of 3 runs
• Code Scanning Fixer — 0 failed, 2 without a verdict, out of 4 runs
• Matt Pocock Skills Reviewer — 0 failed, 1 without a verdict, out of 32 runs

Collected by: https://github.com/github/gh-aw-threat-detection/actions/runs/32689017610
Full data: the detection-stats-32689017610 artifact on that run.

Generated by Detection Stats Daily · auto · 27 AIC · ⌖ 7.84 AIC · ⊞ 11.3K · ◷