[detection-analysis] Detection Analysis Report — 2026-08-16

[github-actions[bot]]

Summary

• Window evaluated: last 24 full hours (UTC), 2026-08-15T23:17:08Z → 2026-08-16T23:17:08Z
• Total runs analyzed: 101
• Detection-enabled runs: 85 (84.2%)
• Regular runs: 16 (15.8%)
• Misconfigured workflows found: 0

Note

No misconfigured workflows were detected against the gh-aw-detection rule set in this window.

Comparison Chart

Group	Runs	Success Rate	Failures	Avg Tokens
Regular	16	75.0%	4	167,109
Detection	85	98.8%	1	59,913

Regular runs are almost entirely intentional CI smoke tests (smoke-* workflows), which explains the lower success rate — several smoke tests exercise failure paths or experimental engines (Gemini, AOAI apikey auth) by design.

Misconfigured Workflows

No misconfigured workflows detected in this window.

Checks performed and results:

1. Frequency rule (gh-aw-detection: false + >3 runs in 7d, using this 24h window as a proxy): no workflow with detection disabled ran more than once in-window.
2. Naming rule (name contains audit/analyzer/report/detector/monitor/inspector without gh-aw-detection: true): all 5 matching workflows in this window (Deep Report, Daily Cache Strategy Analyzer, Agentic Workflow Audit Agent, Daily Regulatory Report Generator, Duplicate Code Detector) already have detection enabled.
3. Detection-step-failure rule: inspected the detection job's steps across all 85 detection-enabled runs — zero failed steps. The one detection-enabled failure (§31976674745, Avenger) failed at "Execute Claude Code CLI", unrelated to detection.
4. Alternating rule: no workflow toggled gh-aw-detection between runs in-window.
5. Documented opt-out for Daily Agentic Workflow AIC Usage Audit (upstream-mirrored) did not need to be applied — it had no runs in this window.

View All Run Metrics

Workflow	Detection	Runs	Success Rate	Avg Tokens
PR Sous Chef	Yes	21	100.0%	38,227
Ponytail Reviewer	Yes	5	100.0%	9,427
Impeccable Skills Reviewer	Yes	5	100.0%	14,929
PR Code Quality Reviewer	Yes	5	100.0%	57,972
Matt Pocock Skills Reviewer	Yes	5	100.0%	21,469
Test Quality Sentinel	Yes	5	100.0%	6,500
Design Decision Gate 🏗️	Yes	5	100.0%	15,818
Issue Monster	Yes	4	100.0%	16,795
Auto-Triage Issues	Yes	2	100.0%	16,316
AI Moderator	Yes	2	100.0%	68,782
PR Description Updater	Yes	2	100.0%	8,268
Avenger	Yes	2	50.0%	22,664
Daily Testify Uber Super Expert	Yes	1	100.0%	10,010
PR Triage Agent	No	1	100.0%	15,815
Copilot Agent PR Analysis	Yes	1	100.0%	22,201
Deep Report	Yes	1	100.0%	69,748
Daily Cache Strategy Analyzer	Yes	1	100.0%	632,581
Daily Code Metrics and Trend Tracking Agent	Yes	1	100.0%	29,060
Daily Safe Output Integrator	Yes	1	100.0%	17,761
[aw] Failure Investigator (6h)	Yes	1	100.0%	46,038
Daily Code Debt Cleanup — Aider	Yes	1	100.0%	5,057
Code Scanning Fixer	Yes	1	100.0%	10,068
Daily Spec Coverage Review	Yes	1	100.0%	8,350
Smoke Copilot Sub Agents	No	1	0.0%	18,990
Daily Graft Intelligence	Yes	1	100.0%	6,706
Daily Ambient Context Optimizer	Yes	1	100.0%	29,133
Daily Team Evolution Insights	No	1	0.0%	13,941
Lockfile Statistics Analysis Agent	No	1	100.0%	13,300
Copilot PR Prompt Pattern Analysis	Yes	1	100.0%	16,358
Daily Project Performance Summary Generator (Using MCP Scripts)	Yes	1	100.0%	22,626
Daily Caveman Optimizer	Yes	1	100.0%	21,820
Contribution Check	Yes	1	100.0%	22,004
Agentic Workflow Audit Agent	Yes	1	100.0%	68,685
Daily Regulatory Report Generator	Yes	1	100.0%	22,262
Smoke Pi	No	1	100.0%	12,328
Smoke Claude	No	1	100.0%	22,485
Agent Container Smoke Test	No	1	100.0%	5,262
Smoke Copilot Auto	No	1	100.0%	3,928
Smoke Codex	No	1	100.0%	564,746
Smoke Gemini	No	1	0.0%	0
Changeset Generator	Yes	1	100.0%	1,083,647
Duplicate Code Detector	Yes	1	100.0%	1,196,933
Smoke Copilot	No	1	100.0%	13,782
Smoke Copilot - AOAI (apikey)	No	1	0.0%	1,491,164
Smoke Copilot - AOAI (Entra)	No	1	100.0%	386,179
Smoke Copilot MAI	No	1	100.0%	90,883
Smoke Copilot Small	No	1	100.0%	4,100
Smoke OTEL	No	1	100.0%	16,846
ESLint Monster	Yes	1	100.0%	18,991
Daily AWF Spec Compiler Surfacing Review	Yes	1	100.0%	0

50 distinct workflows observed across 101 runs.

View Historical Trend

This is the first recorded data point for detection-feature trending; no historical trend chart is available yet. Today's metrics were appended to /tmp/gh-aw/cache-memory/trending/detection/history.jsonl for future comparisons.

Recommendations

• No corrective action needed this cycle — detection coverage is high (84.2% of runs) and all keyword-matching audit/analyzer/report/detector/monitor/inspector workflows already have gh-aw-detection: true.
• Regular (detection-disabled) runs are concentrated in smoke-* CI test workflows, which is expected and intentional — no change recommended there.
• Continue running this report daily to build the historical trend and catch any future drift (e.g., a new audit-style workflow added without detection, or a workflow toggling detection between runs).

References:

• §31976674745 — Avenger (only detection-enabled failure this window; unrelated to detection)
• §31974722434 — Smoke Copilot - AOAI (apikey) (regular-run failure, expected smoke-test behavior)
• §31977783468 — PR Sous Chef (representative detection-enabled run)

Generated by 🔎 Detection Analysis Report · agent · 160 AIC · ⌖ 4.09 AIC · ⊞ 7.7K · ◷

• expires on Aug 19, 2026, 3:27 PM UTC-08:00

[github-actions[bot]]

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

A newer discussion is available at Discussion #53522.