[detection-analysis] Detection Analysis Report — 2026-09-08 (last 24h)

[github-actions[bot]]

Summary

• Window evaluated: last 24 full hours, ~2026-09-07T23:00Z → 2026-09-08T23:00Z (UTC), based on 470 workflow runs pre-downloaded for github/gh-aw
• Total runs analyzed: 470 (331 with complete run metadata; 139 ended before aw_info.json/run_summary.json were produced — see note below)
• Detection-enabled runs: 330 (99.7% of the 331 classifiable runs)
• Regular runs (gh-aw-detection: false): 1 (0.3% of classifiable runs)
• Misconfigured workflows found: 0

Note

No workflows met the misconfiguration criteria in this window (detection explicitly disabled on a high-volume workflow, a failing detection job, or an inconsistent gh-aw-detection setting across runs). The gh-aw-detection feature is running essentially fleet-wide as intended.

On the 139 runs excluded from classification: these runs have only a run.json (no aw_info.json/run_summary.json), meaning the agent job never produced output — mostly scheduled workflows whose activation condition wasn't met (e.g. Daily Trajectory Grader Implementer, [aw] Failure Investigator (6h)) or PR-triggered workflows that failed at pre_activation/activation before the agent step ran (e.g. Front Page Copy Guard, Matt Pocock Skills Reviewer, Ponytail Reviewer). None of these carry features.gh-aw-detection data, so they can't be attributed to either group and are excluded from the comparison below rather than guessed at.

Comparison Chart

With only 1 regular run in the window against 330 detection-enabled runs, the "Regular Runs" side of this comparison is not statistically meaningful this cycle — treat the success-rate/token gap as anecdotal, not a signal that disabling detection is safer.

Misconfigured Workflows

No misconfigured workflows detected in this window.

View All Run Metrics

Showing the 17 workflows with more than one run in the window, plus the single workflow that ran with detection disabled. (127 additional workflows ran exactly once in the window, all with detection enabled by default, and are omitted here for brevity — none of them triggered a misconfiguration rule.)

Workflow	Total Runs	Detection-enabled	Regular	Success Rate	Avg Tokens
PR Sous Chef	70	70	0	95.71%	453,110
Issue Monster	41	41	0	100.0%	59,905
Avenger	13	13	0	0.0%	0
Daily Go Test Parallelizer	10	10	0	100.0%	153,378
Matt Pocock Skills Reviewer	8	8	0	100.0%	21,283
Impeccable Skills Reviewer	8	8	0	100.0%	18,647
Auto-Triage Issues	7	7	0	85.71%	302,995
Test Quality Sentinel	6	6	0	100.0%	8,905
Ponytail Reviewer	6	6	0	100.0%	93,382
PR Code Quality Reviewer	6	6	0	100.0%	677,100
Design Decision Gate 🏗️	6	6	0	100.0%	107,976
AI Moderator	6	6	0	100.0%	448,802
Contribution Check	5	5	0	100.0%	20,144
PR Triage Agent	3	3	0	100.0%	76,396
Deep Report	3	3	0	100.0%	53,885
Code Scanning Fixer	3	3	0	0.0%	14,223
[aw] Failure Investigator (6h)	3	3	0	100.0%	10,625
Daily GitHub Docs SEO Optimizer	1	0	1	100.0%	150,625

View Historical Trend

7 daily snapshots now recorded (2026-08-25 → 2026-09-08). Detection run volume has grown from ~250/day to 330-400/day as more workflows have been added; regular (detection-disabled) run count has stayed near zero throughout, and detection success rate has held in the 75-94% band with no sustained decline.

Recommendations

• No gh-aw-detection misconfigurations require action this cycle — the feature is enabled by default across essentially the entire workflow fleet and no detection job failures were observed.
• Not a detection issue, but worth a separate look: the Avenger workflow failed on all 13 runs in this window (agent job conclusion failure, 0 tokens each run, detection job itself succeeded) — this looks like an agent-level configuration problem unrelated to gh-aw-detection and may warrant its own investigation. Example run: §34183964811.
• Code Scanning Fixer also shows 0% success across its 3 runs this window; same caveat — agent-level, not a detection misconfiguration, but worth a look. Example run: §34183946323.
• Keep monitoring the 139 runs/day that end before aw_info.json is produced — most look like expected no-op activations, but a recurring pre_activation/activation failure pattern (e.g. Front Page Copy Guard, Design Decision Gate) would be worth a dedicated audit outside this detection-focused report.

References:

• §34181032740 — Matt Pocock Skills Reviewer (sample detection-enabled run)
• §34183964811 — Avenger (sample agent-failure run, unrelated to detection)
• §34188208502 — Daily GitHub Docs SEO Optimizer (the one regular/detection-disabled run this window)

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• api.anthropic.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "api.anthropic.com"

See Network Configuration for more information.

Generated by 🔎 Detection Analysis Report · claude · agent · 219.1 AIC · ⌖ 10.5 AIC · ⊞ 9K · ◷

• expires on Sep 11, 2026, 3:39 PM UTC-08:00

[github-actions[bot]]

🤖 Beep boop! Your friendly neighborhood smoke test bot swooped by to say hi. If code review had a mascot, it would be a very caffeinated robot cat chasing a laser pointer made of green checkmarks. Stay awesome, gh-aw community! 🐱✅

Warning

Firewall blocked 7 domains

The following domains were blocked by the firewall during workflow execution:

• accounts.google.com
• android.clients.google.com
• clients2.google.com
• contentautofill.googleapis.com
• host.docker.internal
• www.google.com
• www.gstatic.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "accounts.google.com"
    - "android.clients.google.com"
    - "clients2.google.com"
    - "contentautofill.googleapis.com"
    - "host.docker.internal"
    - "www.google.com"
    - "www.gstatic.com"

See Network Configuration for more information.

📰 BREAKING: Report filed by Smoke Copilot · copilot · auto · 39.3 AIC · ⌖ 2.56 AIC · ⊞ 10K · ◷
_{Comment /smoke-copilot to run again
Add label smoke to run again}

[github-actions[bot]]

This discussion has been marked as outdated by Detection Analysis Report.

A newer discussion is available at Discussion #59818.