[detection-analysis] Detection Analysis Report — 2026-08-30

[github-actions[bot]]

Summary

• Window evaluated: last 24 full hours (UTC), 2026-08-29T23:41:03Z → 2026-08-30T23:41:03Z
• Total runs analyzed: 406 (across 130 distinct workflows)
• Detection-enabled runs: 403 (99.3%)
• Regular runs (gh-aw-detection: false): 3 (0.7%)
• Misconfigured workflows found: 3

Warning

3 workflows explicitly disable the gh-aw-detection feature while running frequently enough (>3 runs in the last 7 days) that detection would meaningfully help catch anomalies. No detection-job-level failures or flag-alternation issues were found.

Comparison Chart

Detection-enabled runs vastly outnumber regular runs (403 vs 3) — the run-count panel uses a log scale to keep the "Regular" bar visible. Detection-enabled runs also show a higher success rate (87.6% vs 66.7%) and consume ~6x more tokens on average (129.9k vs 21.0k), consistent with the extra detection-analysis step running as part of those workflows.

Misconfigured Workflows

Workflow	Misconfiguration Type	7-day Run Count	Recommended Fix
Smoke Codex	gh-aw-detection: false on a high-frequency workflow (>3 runs/7d)	14	Remove the explicit gh-aw-detection: false override (or justify it in a comment) so detection runs by default and can catch anomalous smoke-test behavior.
Smoke Copilot	gh-aw-detection: false on a high-frequency workflow (>3 runs/7d)	21	Same as above — re-enable detection given its high run cadence.
Daily GitHub Docs SEO Optimizer	gh-aw-detection: false on a high-frequency workflow (>3 runs/7d)	8	Same as above — re-enable detection; this workflow also has a 0% success rate in-window and would benefit from detection-assisted diagnosis.

Example runs (24h window): §33287374932 (Smoke Codex), §33288730255 (Smoke Copilot), §33293189773 (Daily GitHub Docs SEO Optimizer).

No workflows were found with a failing detection job conclusion (rule 2), and no workflow alternated between detection-enabled and detection-disabled within the window (rule 3).

View All Run Metrics

Workflows with ≥2 runs in the window (20 of 130 distinct workflows; the remaining 110 each had exactly 1 run):

Workflow	Detection	Runs	Success %	Avg Tokens	Example Run
PR Sous Chef	Enabled (explicit)	83	100.0%	148,346	§33282107607
Issue Monster	Enabled (explicit)	44	100.0%	104,872	§33284574563
Daily Trajectory Grader Implementer	Enabled (explicit)	36	100.0%	10,925	§33284913251
Test Quality Sentinel	Enabled (explicit)	15	100.0%	8,214	§33281820639
PR Code Quality Reviewer	Enabled (explicit)	14	100.0%	553,865	§33281820612
Design Decision Gate 🏗️	Enabled (explicit)	14	100.0%	58,558	§33281820626
Impeccable Skills Reviewer	Enabled (explicit)	14	100.0%	7,721	§33281820630
Matt Pocock Skills Reviewer	Enabled (explicit)	14	100.0%	12,966	§33281820715
Ponytail Reviewer	Enabled (explicit)	14	100.0%	207,713	§33281820731
Daily Go Test Parallelizer	Enabled (explicit)	11	0.0%	0	§33284910807
Contribution Check	Enabled (explicit)	6	100.0%	14,065	§33285175552
Auto-Triage Issues	Enabled (explicit)	5	100.0%	59,401	§33284654255
PR Triage Agent	Enabled (default)	4	100.0%	69,529	§33284620398
Deep Report	Enabled (explicit)	4	75.0%	52,978	§33284637093
[aw] Failure Investigator (6h)	Enabled (explicit)	4	75.0%	9,449	§33285148967
Code Scanning Fixer	Enabled (explicit)	4	0.0%	11,395	§33285210521
PR Description Updater	Enabled (explicit)	3	100.0%	12,799	§33291539231
Avenger	Enabled (explicit)	3	0.0%	0	§33294778348
Daily Credit Limit Test	Enabled (explicit)	2	100.0%	125,131	§33285204348
AI Moderator	Enabled (explicit)	2	100.0%	48,558	§33313387886

"Enabled (default)" = gh-aw-detection key absent from aw_info.json (the feature default). "Enabled (explicit)" = key present and true. "Disabled" = key present and false.

Aggregate metrics:

• Regular runs: 3, success rate 66.67%, avg tokens 21,013.67
• Detection runs: 403, success rate 87.59%, avg tokens 129,860.12

View Historical Trend

Insufficient history for a trend chart — the cache-memory trending store (trending/detection-metrics/history.jsonl) has 5 daily entries after today's append; a 30-day trend chart is generated once ≥7 entries exist.

Recent daily snapshots:

Date (UTC)	Regular Runs	Detection Runs	Regular Success %	Detection Success %	Misconfigured
2026-08-25	0	251	0.0%	92.03%	0
2026-08-28 (early)	2	286	100.0%	75.87%	0
2026-08-28 (late)	3	348	66.67%	91.67%	0
2026-08-29	11	319	90.9%	94.0%	0
2026-08-30 (today)	3	403	66.67%	87.59%	3

Today is the first day this window's rules flagged any misconfigured workflows — worth watching whether Smoke Codex / Smoke Copilot / Daily GitHub Docs SEO Optimizer keep the gh-aw-detection: false override going forward.

Recommendations

1. Re-enable (or justify) gh-aw-detection on Smoke Codex, Smoke Copilot, and Daily GitHub Docs SEO Optimizer — all three run frequently (8-21 times/week) with the feature explicitly off, forgoing anomaly detection on a meaningful share of CI volume.
2. Investigate Daily GitHub Docs SEO Optimizer's 0% success rate in this window (run §33293189773, 0 tokens recorded) — it looks like the agent job didn't execute; re-enabling detection would help diagnose future failures like this one.
3. Daily Go Test Parallelizer (11 runs, 0% success) and Code Scanning Fixer (4 runs, 0% success) are worth a separate look — detection is already enabled for both, so this is an agent-level reliability issue rather than a detection misconfiguration.
4. Continue accumulating daily trend snapshots — 2 more days of data will unlock the 30-day trend chart.

References:

• §33287374932 — Smoke Codex (detection disabled)
• §33288730255 — Smoke Copilot (detection disabled)
• §33293189773 — Daily GitHub Docs SEO Optimizer (detection disabled, failed run)

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 · 303.5 AIC · ⌖ 8.38 AIC · ⊞ 8.8K · ◷

• expires on Sep 2, 2026, 3:54 PM UTC-08:00

[github-actions[bot]]

🤖 Smoke test bot checking in on this Detection Analysis discussion! Everything looks in order here — carry on with the great analysis work. 🎉

Warning

Firewall blocked 6 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
• 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"
    - "www.google.com"
    - "www.gstatic.com"

See Network Configuration for more information.

📰 BREAKING: Report filed by Smoke Copilot · copilot · auto · 33.5 AIC · ⌖ 2.49 AIC · ⊞ 9.1K · ◷
_{Comment /smoke-copilot to run again
Add label smoke to run again}