[detection-analysis] Detection Analysis Report — 2026-07-18 (2 misconfigured workflows)

[github-actions[bot]]

Summary

Warning

2 misconfigured workflows were detected in this window. See details below.

• Window evaluated: last 24 full hours (UTC) — 2026-07-17T13:18:01Z → 2026-07-18T13:18:01Z (observed runs span 2026-07-18T05:46Z → 2026-07-18T12:55Z)
• Total runs analyzed: 162
• Detection-enabled runs (gh-aw-detection: true): 37 (22.8%)
• Regular runs (false / absent): 125 (77.2%)
• Misconfigured workflows found: 2

Metric	Regular runs	Detection runs
Run count	125	37
Success rate	88.0%	94.6%
Failures	15	2
Avg tokens	68,755	82,748

Detection-enabled runs show a +6.6 pp higher success rate despite higher average token usage.

Comparison Chart

Misconfigured Workflows

Workflow	Misconfiguration	Runs	Recommended fix
Duplicate Code Detector	Name implies monitoring/analysis (detector) but gh-aw-detection is not enabled	1	Add gh-aw-detection: true to the workflow frontmatter features.
Smoke CI	gh-aw-detection: false on a frequently-run workflow (>3 runs in window)	9	Set gh-aw-detection: true — unless detection is intentionally disabled (see note).

Note

Smoke CI — all 9 runs report 0 tokens and a non-success conclusion, consistent with a non-agentic lock-compilation / CI smoke job rather than an agentic run. If this workflow does not invoke an agent, detection is not applicable and the flag can be treated as an intentional opt-out. Recommend confirming and, if so, documenting the opt-out to suppress future flags.

Daily Agentic Workflow AIC Usage Audit was not flagged (documented upstream opt-out); it did not run in this window.

View All Run Metrics

Detection column: ✅ on = all runs gh-aw-detection: true; — off = false/absent; ⚠️ mixed = alternates within window.

Workflow	Detection	Runs	Success %	Avg tokens
Design Decision Gate 🏗️	— off	12	100.0	10,928
Impeccable Skills Reviewer	— off	12	100.0	8,746
Matt Pocock Skills Reviewer	— off	12	100.0	15,486
PR Code Quality Reviewer	— off	12	83.3	25,292
Test Quality Sentinel	✅ on	12	100.0	8,316
PR Sous Chef	— off	11	90.9	42,019
Smoke CI	— off	9	0.0	0
Issue Monster	— off	8	100.0	21,940
Tidy	— off	8	75.0	5,668
PR Description Updater	— off	6	100.0	12,546
Auto-Triage Issues	✅ on	2	100.0	10,234
PR Triage Agent	— off	2	100.0	12,852
Agent Container Smoke Test	— off	1	100.0	4,117
Agent Performance Analyzer - Meta-Orchestrator	✅ on	1	100.0	15,823
CLI Version Checker	✅ on	1	100.0	70,299
Changeset Generator	✅ on	1	100.0	0
Claude Code User Documentation Review	✅ on	1	100.0	67,183
Code Scanning Fixer	✅ on	1	100.0	9,490
Constraint Solving — Problem of the Day	✅ on	1	100.0	8,720
Contribution Check	✅ on	1	100.0	18,864
Copilot Agent Prompt Clustering Analysis	✅ on	1	100.0	39,178
Copilot Session Insights	✅ on	1	100.0	50,570
Daily Assign Issue To User	✅ on	1	100.0	3,123
Daily Cli Tools Tester	✅ on	1	100.0	17,343
Daily Documentation Updater	✅ on	1	100.0	79,696
Daily Evals Feature Report	✅ on	1	100.0	252,772
Daily Go Function Namer	✅ on	1	100.0	17,962
Daily Hippo Learn	✅ on	1	100.0	56,224
Daily Max Ai Credits Test	✅ on	1	0.0	3,011
Daily Rendering Scripts Verifier	✅ on	1	100.0	56,279
Daily Safe Outputs Conformance Checker	— off	1	100.0	15,281
Daily Sub-Agent Model Resolution Audit	✅ on	1	100.0	1,910,789
Daily Syntax Error Quality Check	— off	1	100.0	8,089
Daily Workflow Updater	— off	1	100.0	4,297
Dev	— off	1	100.0	505,911
Discussion Task Miner - Code Quality Improvement Agent	— off	1	100.0	15,948
Draft PR Cleanup	— off	1	100.0	5,484
Duplicate Code Detector	— off	1	100.0	137,111
ESLint Miner	— off	1	100.0	23,645
ESLint Refiner	— off	1	100.0	71,827
GitHub API Consumption Report Agent	✅ on	1	100.0	39,922
GitHub Remote MCP Authentication Test	✅ on	1	100.0	24,554
Instructions Janitor	— off	1	100.0	65,143
Issue Arborist	— off	1	100.0	705,982
Multi-Device Docs Tester	✅ on	1	100.0	8,100
Package Specification Enforcer	— off	1	100.0	44,768
Package Specification Extractor	— off	1	100.0	6,936
Release	— off	1	100.0	9,697
Schema Consistency Checker	— off	1	100.0	46,250
Smoke Antigravity	— off	1	100.0	0
Smoke Claude	— off	1	100.0	22,827
Smoke Codex	— off	1	100.0	267,887
Smoke Copilot	— off	1	0.0	2,561,246
Smoke Copilot - AOAI (Entra)	— off	1	100.0	369,784
Smoke Copilot - AOAI (apikey)	— off	1	100.0	463,770
Smoke Copilot MAI	— off	1	100.0	183,915
Smoke Copilot Small	— off	1	100.0	3,575
Smoke Gemini	— off	1	100.0	332,661
Smoke OTEL	— off	1	100.0	14,292
Smoke Pi	— off	1	100.0	13,395
Stale PR Cleanup	— off	1	100.0	6,859
Sub-Issue Closer	— off	1	100.0	10,390
Team Status	— off	1	100.0	36,016
Terminal Stylist	— off	1	100.0	93,937
Visual Regression Checker	— off	1	100.0	10,020
Weekly Editors Health Check	— off	1	100.0	20,817
[aw] Failure Investigator (6h)	✅ on	1	100.0	26,548
daily-experiment-report	✅ on	1	0.0	82,199

View Historical Trend

30-day trend of regular vs. detection run volume, with detection success-rate band (secondary axis). Today's metrics were appended to the cache-memory trending store (21 total daily entries).

Recommendations

1. Duplicate Code Detector — enable gh-aw-detection: true. Its analysis-oriented purpose is exactly the intended detection use case; it currently runs without it.
2. Smoke CI — confirm whether this is an agentic workflow. If it is purely a CI/lock smoke test (its 0-token, non-success signature suggests so), document the opt-out so it stops being flagged. If it is agentic, enable detection.
3. Detection failures — 2 detection-enabled runs failed: daily-experiment-report (§29637424828) and Daily Max Ai Credits Test (§29642635802). Review their logs; these are the only detection-side reliability gaps this window.
4. Coverage — only 22.8% of runs have detection enabled. Consider enabling it on the remaining audit/report/analyzer-class workflows to widen monitoring coverage.

Note

The >3 runs misconfiguration rule nominally spans 7 days; this run only had ~7 hours of log data available (the logs fetch returned runs from 05:46–12:55 UTC), so run-count thresholds are evaluated over the observed window and may undercount.

References:

• §29645485336 (this analysis run)
• §29637424828 (daily-experiment-report failure)
• §29642635802 (Daily Max Ai Credits Test failure)

Warning

Firewall blocked 1 domain

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

• awmgmcpg

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

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by 🔎 Detection Analysis Report · 195.2 AIC · ⌖ 39.9 AIC · ⊞ 6.9K · ◷

• expires on Jul 21, 2026, 5:25 AM UTC-08:00

[github-actions[bot]]

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

A newer discussion is available at Discussion #46686.