522 external-detector runs analysed · detection-job error rate 0% · soft failures 26 · threat rate 0.90% (4/444 runs with a verdict)

Summary

gh-aw detection statistics - 2026-08-18 (UTC)

Repository: github/gh-aw
Window: 2026-08-18T00:00:00Z .. 2026-08-18T23:59:59Z
API requests: 2980, rate-limit pauses: 1
Data complete: yes

Totals

Metric	Count
Workflow runs in window	3569
Agentic runs (`*.lock.yml`)	2022
Runs with a `detection` job	1271
... using the external detector	522
... using the built-in detector	89
... detector could not be determined	660
Agentic runs without a `detection` job	751

All rates below are over the external detector population (522 runs). A run is external when its detection job showed the Install threat-detect binary step, or when another run of the same workflow did that day — a job that was skipped or died during setup reports no steps and would otherwise be miscounted.

Detection job outcomes

Outcome	Count	Rate
`success`	470	90.04%
`skipped`	52	9.96%

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

Verdict availability

State	Meaning	Count
`present`	detection artifact downloaded and parsed	444
`absent`	detection job ran but published no artifact (soft failure)	26
`skipped`	detection job was skipped or was still running (nothing to fetch)	52

Green detection jobs that published no verdict: 26 (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	444
Clean (no threat)	440
Any threat	4
`prompt_injection`	1
`secret_leak`	1
`malicious_patch`	2

Threat rate (of runs with a verdict): 0.9%

Reasons reported by gh-aw

From the [aw] Detection Runs tracking issue (warning/failure conclusions only).

Reason	Count
`agent_failure`	31
`threat_detected`	7
`parse_error`	1

By workflow

Workflow	Runs	Skipped	No verdict	Threats
PR Sous Chef	79	0	0	1
Issue Monster	47	27	0	1
Test Quality Sentinel	40	0	0	0
Matt Pocock Skills Reviewer	35	0	0	0
Impeccable Skills Reviewer	34	0	0	0
PR Code Quality Reviewer	34	0	0	0
Ponytail Reviewer	34	0	0	0
Design Decision Gate 🏗️	33	0	2	0
PR Description Updater	24	7	0	2
Avenger	19	8	1	0
Auto-Triage Issues	16	8	0	0
Daily Go Test Parallelizer	12	2	0	0
Contribution Check	6	0	0	0
Code Scanning Fixer	4	0	0	0
Deep Report	4	0	0	0
[aw] Failure Investigator (6h)	4	0	0	0
Daily Credit Limit Test	2	0	1	0
Agent Job Health Monitor	1	0	0	0
Agent Performance Analyzer - Meta-Orchestrator	1	0	0	0
Agent Persona Explorer	1	0	0	0
Agentic Workflow Audit Agent	1	0	1	0
Architecture Guardian	1	0	1	0
CI Optimization Coach	1	0	1	0
CLI Consistency Checker	1	0	1	0
CLI Version Checker	1	0	1	0

87 further workflows omitted; see stats.json.

Notable runs

Workflow	Run	Job conclusion	Verdict	Threats	Reported reason
Issue Monster	32142125210	`success`	`present`	prompt_injection	`threat_detected`
PR Description Updater	32095718588	`success`	`present`	malicious_patch	`threat_detected`
PR Description Updater	32101549737	`success`	`present`	secret_leak	`threat_detected`
PR Sous Chef	32109331559	`success`	`present`	malicious_patch	`threat_detected`
Agentic Workflow Audit Agent	32186551970	`success`	`absent`	-	`agent_failure`
Architecture Guardian	32147417635	`success`	`absent`	-	`agent_failure`
Avenger	32155564962	`success`	`absent`	-	`agent_failure`
CI Optimization Coach	32142141919	`success`	`absent`	-	`agent_failure`
CLI Consistency Checker	32142179224	`success`	`absent`	-	`agent_failure`
CLI Version Checker	32103280568	`success`	`absent`	-	`agent_failure`
Cache directory setup	32197746932	`success`	`absent`	-	`agent_failure`
Copilot Agent PR Analysis	32170153054	`success`	`absent`	-	`-`
Daily AIC Consumption Report (Sentry + Grafana OTel)	32134250479	`success`	`absent`	-	`agent_failure`
Daily AWF Spec Compiler Surfacing Review	32191955239	`success`	`absent`	-	`agent_failure`
Daily BYOK Ollama Test	32192747973	`success`	`absent`	-	`agent_failure`
Daily Code Debt Cleanup — Aider	32174369814	`success`	`absent`	-	`agent_failure`
Daily Code Metrics and Trend Tracking Agent	32171669273	`success`	`absent`	-	`agent_failure`
Daily Credit Limit Test	32087739571	`success`	`absent`	-	`agent_failure`
Daily Documentation Updater	32128281831	`success`	`absent`	-	`agent_failure`
Daily Fact	32147275242	`success`	`absent`	-	`agent_failure`
Daily Go Test Stubs — Aider	32161145619	`success`	`absent`	-	`agent_failure`
Daily Max Ai Credits Test	32130849835	`success`	`absent`	-	`agent_failure`
Daily Reliability Review	32195004998	`success`	`absent`	-	`agent_failure`
Daily Sub-Agent Model Resolution Audit	32103509594	`success`	`absent`	-	`agent_failure`
Daily Team Evolution Insights	32180976336	`success`	`absent`	-	`agent_failure`
Design Decision Gate 🏗️	32089118620	`success`	`absent`	-	`agent_failure`
Design Decision Gate 🏗️	32101513747	`success`	`absent`	-	`agent_failure`
Duplicate Code Detector	32189950672	`success`	`absent`	-	`agent_failure`
arXiv Paper Researcher: GitHub Agentic Workflows	32117088930	`success`	`absent`	-	`agent_failure`
daily-experiment-report	32114181360	`success`	`absent`	-	`agent_failure`

Change since 2026-08-17

Metric	2026-08-17	2026-08-18	Delta
External detector runs	555	522	-33
Error rate	0.36%	0.00%	-0.36 pp
Soft failures	29	26	-3
Runs with a verdict	377	444	+67
Any threat	1	4	+3
Threat rate	0.27%	0.90%	+0.63 pp

Only 2 history entries available; 7-day mean will appear once 7 days have accumulated.

Watch list

Design Decision Gate 🏗️ — 0 failed, 2 without a verdict, out of 33 runs
Avenger — 0 failed, 1 without a verdict, out of 19 runs
Daily Credit Limit Test — 0 failed, 1 without a verdict, out of 2 runs
Agentic Workflow Audit Agent — 0 failed, 1 without a verdict, out of 1 runs
Architecture Guardian — 0 failed, 1 without a verdict, out of 1 runs

Collected by: https://github.com/github/gh-aw-threat-detection/actions/runs/32214573600
Full data: the detection-stats-32214573600 artifact on that run.

Generated by Detection Stats Daily · 27.2 AIC · ⌖ 16.1 AIC · ⊞ 7.2K · ◷