[detection-analysis] Detection Analysis Report — 2026-07-21 (last 24h)

[github-actions[bot]]

Summary

• Window evaluated: last 24 full hours (UTC) — 2026-07-20T23:37:05Z to 2026-07-21T23:37:05Z
• Total runs analyzed: 261
• Detection-enabled runs: 88 (33.7%)
• Regular runs: 173 (66.3%)
• Misconfigured workflows found: 3

Warning

3 misconfigured workflows were detected in this window — see the section below.

Metric	Regular Runs	Detection Runs
Run count	173	88
Success rate	96.5%	92.0%
Failures	6	7
Avg tokens / run	36,218	69,056

Comparison Chart

Misconfigured Workflows

Workflow	Misconfiguration Type	Run Count	Example Run	Recommended Fix
Deep Report	name-suggests-detection-missing	1	§29843150600	Add gh-aw-detection: true to the workflow, or document an explicit opt-out.
Daily Security Observability Report	name-suggests-detection-missing	1	§29848004889	Add gh-aw-detection: true to the workflow, or document an explicit opt-out.
Daily Regulatory Report Generator	name-suggests-detection-missing	1	§29870516386	Add gh-aw-detection: true to the workflow, or document an explicit opt-out.

All three flags are rule-2 (name-based) matches: the workflow name contains an analysis keyword (report) but gh-aw-detection: true is not set. The documented opt-out Daily Agentic Workflow AIC Usage Audit was excluded from flagging. No rule-1 (high-volume detection-disabled), rule-3 (detection step failed), or rule-4 (alternating) matches were found — the 12 explicit gh-aw-detection: false runs are all single-run smoke tests, and no workflow alternated its detection setting.

View All Run Metrics

Per-workflow breakdown for all 118 distinct workflows in the window (58 detection-enabled, 60 regular, 0 alternating).

Workflow	Runs	Detection	Success %	Avg Tokens	Engine
PR Sous Chef	24	no	100.0	47,786	pi
Matt Pocock Skills Reviewer	20	no	100.0	13,715	copilot
Design Decision Gate 🏗️	18	no	94.4	10,371	claude
Impeccable Skills Reviewer	17	no	100.0	7,149	copilot
PR Code Quality Reviewer	17	no	94.1	18,940	copilot
Test Quality Sentinel	17	yes	94.1	10,026	copilot
Tidy	8	no	75.0	4,250	copilot
Issue Monster	7	no	100.0	17,588	pi
PR Description Updater	6	no	100.0	8,192	copilot
AI Moderator	5	yes	100.0	58,167	codex
Auto-Triage Issues	4	yes	100.0	16,122	pi
Contribution Check	4	yes	100.0	24,718	copilot
Code Scanning Fixer	3	yes	100.0	4,986	copilot
Discussion Task Miner - Code Quality Improvement Agent	3	no	100.0	15,279	copilot
PR Triage Agent	3	no	100.0	15,041	copilot
[aw] Failure Investigator (6h)	3	yes	100.0	27,833	claude
Agent Container Smoke Test	1	no	100.0	4,404	copilot
Agent Performance Analyzer - Meta-Orchestrator	1	yes	100.0	22,330	copilot
Agent Persona Explorer	1	yes	100.0	52,906	pi
Agentic Workflow AIC Usage Optimizer	1	no	100.0	17,546	copilot
Agentic Workflow Audit Agent	1	yes	100.0	46,691	claude
Architecture Guardian	1	yes	100.0	4,106	copilot
Avenger	1	yes	100.0	0	claude
Breaking Change Checker	1	yes	0.0	8,831	copilot
CI Optimization Coach	1	yes	0.0	0	copilot
CLI Consistency Checker	1	yes	100.0	67,516	copilot
Changeset Generator	1	yes	100.0	67,198	codex
Claude Code User Documentation Review	1	yes	100.0	86,407	claude
Constraint Solving — Problem of the Day	1	yes	100.0	8,715	copilot
Copilot Agent PR Analysis	1	yes	100.0	14,867	claude
Copilot Agent Prompt Clustering Analysis	1	yes	100.0	37,091	claude
Copilot PR Conversation NLP Analysis	1	yes	100.0	41,870	copilot
Copilot PR Prompt Pattern Analysis	1	yes	100.0	12,929	copilot
Copilot Session Insights	1	yes	100.0	53,100	claude
Daily AIC Consumption Report (Sentry + Grafana OTel)	1	yes	100.0	41,079	claude
Daily AW Cross-Repo Compile Check	1	yes	100.0	26,831	claude
Daily AWF Spec Compiler Surfacing Review	1	yes	100.0	54,117	pi
Daily Agent of the Day Blog Writer	1	yes	100.0	48,962	copilot
Daily Agentic Workflow AIC Usage Audit	1	no	100.0	11,011	copilot
Daily Ambient Context Optimizer	1	yes	100.0	38,668	copilot
Daily BYOK Ollama Test	1	yes	0.0	0	copilot
Daily CLI Performance Agent	1	yes	100.0	16,774	copilot
Daily Cache Strategy Analyzer	1	yes	100.0	160,268	codex
Daily Caveman Optimizer	1	yes	100.0	26,085	claude
Daily Choice Type Test	1	yes	100.0	3,949	claude
Daily Code Metrics and Trend Tracking Agent	1	yes	100.0	32,702	claude
Daily Copilot PR Merged Report	1	yes	100.0	11,456	copilot
Daily Documentation Updater	1	yes	100.0	63,341	pi
Daily Elixir Credo Snippet Audit	1	yes	100.0	7,902	claude
Daily Evals Feature Report	1	yes	100.0	2,620,068	codex
Daily Fact	1	yes	100.0	18,376	codex
Daily File Diet	1	yes	100.0	13,787	copilot
Daily Formal Spec Verifier	1	yes	0.0	8,194	copilot
Daily Go Function Namer	1	yes	100.0	15,425	pi
Daily Hippo Learn	1	yes	100.0	84,135	pi
Daily Issues Report Generator	1	yes	0.0	11,931	copilot
Daily MCP Tool Concurrency Analysis	1	yes	100.0	10,958	copilot
Daily Malicious Code Scan Agent	1	yes	100.0	9,418	copilot
Daily Max Ai Credits Test	1	yes	0.0	2,972	copilot
Daily News	1	yes	100.0	75,310	pi
Daily Project Performance Summary Generator (Using MCP Scripts)	1	yes	100.0	16,154	copilot
Daily Regulatory Report Generator	1	no	100.0	7,152	copilot
Daily Reliability Review	1	yes	100.0	25,961	claude
Daily Rendering Scripts Verifier	1	yes	100.0	41,029	claude
Daily SPDD Spec Planner	1	no	100.0	9,310	copilot
Daily Safe Output Integrator	1	no	100.0	8,804	copilot
Daily Secrets Analysis Agent	1	no	100.0	9,338	copilot
Daily Security Observability Report	1	no	100.0	15,337	copilot
Daily Syntax Error Quality Check	1	no	100.0	7,739	copilot
Daily Team Evolution Insights	1	no	100.0	18,814	claude
Daily Testify Uber Super Expert	1	yes	100.0	16,566	copilot
Daily VulnHunter Scan	1	no	0.0	88,521	claude
Daily Workflow Updater	1	no	100.0	4,366	copilot
Dead Code Removal Agent	1	no	100.0	19,202	copilot
Deep Report	1	no	100.0	30,832	claude
DeepSec Security Scan	1	no	100.0	5,739	claude
Delight	1	no	100.0	13,521	copilot
Dev	1	no	100.0	282,977	codex
Draft PR Cleanup	1	no	100.0	5,697	copilot
Duplicate Code Detector	1	yes	100.0	1,112,766	codex
ESLint Miner	1	no	100.0	33,506	copilot
ESLint Monster	1	no	100.0	16,372	pi
Functional Pragmatist	1	no	100.0	10,774	copilot
GEO Optimizer Daily Audit	1	yes	100.0	11,497	copilot
GitHub API Consumption Report Agent	1	yes	100.0	39,439	claude
GitHub MCP Structural Analysis	1	yes	100.0	28,252	claude
Glossary Maintainer	1	no	100.0	16,856	copilot
Go Fan	1	no	100.0	28,368	claude
Instructions Janitor	1	no	100.0	82,210	claude
Issue Triage Agent	1	no	100.0	26,672	copilot
Linter Miner	1	no	100.0	16,981	copilot
Lockfile Statistics Analysis Agent	1	no	100.0	16,102	claude
Package Specification Enforcer	1	no	100.0	35,807	pi
Package Specification Extractor	1	no	100.0	14,488	copilot
Release	1	no	100.0	13,385	copilot
Repository Quality Improvement Agent	1	no	100.0	22,121	copilot
Slide Deck Maintainer	1	no	100.0	7,555	copilot
Smoke Antigravity	1	no	100.0	0	antigravity
Smoke Claude	1	no	100.0	26,821	claude
Smoke Claude on Copilot	1	no	0.0	0	claude
Smoke Codex	1	no	100.0	108,562	codex
Smoke Copilot	1	no	100.0	766,472	copilot
Smoke Copilot - AOAI (Entra)	1	no	100.0	570,285	copilot
Smoke Copilot - AOAI (apikey)	1	no	100.0	983,613	copilot
Smoke Copilot MAI	1	no	100.0	77,724	copilot
Smoke Copilot Small	1	no	100.0	3,203	copilot
Smoke Copilot Sub Agents	1	no	100.0	33,290	copilot
Smoke Gemini	1	no	100.0	206,010	gemini
Smoke OTEL	1	no	100.0	14,178	copilot
Smoke Pi	1	no	100.0	13,820	pi
Sub-Issue Closer	1	no	100.0	11,010	copilot
Team Status	1	no	100.0	70,076	copilot
Terminal Stylist	1	no	100.0	97,293	copilot
The Daily Repository Chronicle	1	yes	100.0	14,143	copilot
Typist - Go Type Analysis	1	yes	100.0	94,887	claude
UK AI Operational Resilience	1	no	100.0	21,888	copilot
Workflow Normalizer	1	no	100.0	11,409	copilot
daily-experiment-report	1	yes	100.0	25,909	copilot

View Historical Trend

30-day trend of regular vs. detection run volume, with the detection success-rate band (23 daily data points in cache-memory history).

Recommendations

1. Enable detection on the three flagged report workflows (Deep Report, Daily Security Observability Report, Daily Regulatory Report Generator) by adding gh-aw-detection: true to their frontmatter — or, if any is an intentional opt-out, document it in the repository-level opt-out list the way Daily Agentic Workflow AIC Usage Audit is documented.
2. Detection success rate (92.0%) trails regular runs (96.5%) this window; the 7 detection failures are worth a spot-check, though the 30-day trend shows detection success consistently in the high-80s/low-90s and stable.
3. Detection runs cost ~1.9× the tokens of regular runs (69,056 vs 36,218 avg) — expected given the extra threat-detection pass; no action needed, tracked for cost awareness.
4. Detection adoption is healthy at 33.7% of runs and the volume trend is rising alongside overall run growth.

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 · age00 234.6 AIC · ⌖ 40.6 AIC · ⊞ 6.8K · ◷

• expires on Jul 24, 2026, 3:51 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 #47427.