Description
Several audit/report-class workflows that aggregate cross-run agentic output are running without threat detection on their outputs. Verified missing gh-aw-detection in frontmatter on main:
.github/workflows/agentic-token-audit.md (Daily Agentic Workflow AIC Usage Audit)
.github/workflows/daily-token-consumption-report.md (Daily AIC Consumption Report)
.github/workflows/agentic-token-trend-audit.md
.github/workflows/agentic-token-optimizer.md
Audit/report workflows ingest and summarize other runs' outputs, so they should scan those aggregated outputs. Fix is a one-line gh-aw-detection: true in each frontmatter, then recompile the lock files.
Expected Impact
Closes a threat-detection coverage gap on the highest-aggregation workflows; the detection report shows detection-enabled runs also had a higher success rate (91.7% vs 76.5%) this window.
Suggested Agent
Agentic Maintenance / Copilot SWE Agent.
Estimated Effort
Fast (< 30 min) — one frontmatter line each + recompile.
Data Source
DeepReport 2026-06-30; detection-analysis report #42485.
Generated by 🔬 Deep Report · 204.7 AIC · ⌖ 16.9 AIC · ⊞ 10.2K · ◷
Description
Several audit/report-class workflows that aggregate cross-run agentic output are running without threat detection on their outputs. Verified missing
gh-aw-detectionin frontmatter onmain:.github/workflows/agentic-token-audit.md(Daily Agentic Workflow AIC Usage Audit).github/workflows/daily-token-consumption-report.md(Daily AIC Consumption Report).github/workflows/agentic-token-trend-audit.md.github/workflows/agentic-token-optimizer.mdAudit/report workflows ingest and summarize other runs' outputs, so they should scan those aggregated outputs. Fix is a one-line
gh-aw-detection: truein each frontmatter, then recompile the lock files.Expected Impact
Closes a threat-detection coverage gap on the highest-aggregation workflows; the detection report shows detection-enabled runs also had a higher success rate (91.7% vs 76.5%) this window.
Suggested Agent
Agentic Maintenance / Copilot SWE Agent.
Estimated Effort
Fast (< 30 min) — one frontmatter line each + recompile.
Data Source
DeepReport 2026-06-30; detection-analysis report #42485.