[optimization:ai-credit-auditor] AI Credit Audit — 2026-08-28 (953.56 AIC / $9.54 est.)

📊 Executive Summary

• Period: last 24 hours (2026-08-27 to 2026-08-28)
• Total runs: 309 (all completed)
• Total AI credits: 953.5553 AIC
• Estimated observed cost: $9.5356 USD (total AI credits * $0.01; billing estimate — dashboards are authoritative)
• Total tokens: 1,157,851
• Total Actions minutes: 322.0 min
• Active workflows: 18
• Next 7 days: P10 30.75 AIC / $0.31, P50 266.08 AIC / $2.66, P90 649.27 AIC / $6.49
• Next 30 days: P10 635.83 AIC / $6.36, P50 1285.55 AIC / $12.86, P90 2205.24 AIC / $22.05

🏆 Top 5 Workflows by AI Credit Spend

Cost Forecast

Forecasts are estimates derived from historical samples (30-day monte-carlo simulation, gh aw forecast, exit_code=0, valid JSON). Billing dashboards are authoritative.

Repository-level totals above are sums of per-workflow projections, not an independently simulated portfolio confidence interval.

Top workflows by monthly P50 AIC:

Forecast Confidence and Data Quality

• Forecast command completed successfully: exit_code=0, json_valid=true.
• 54 data-quality flags were raised across 40 tracked workflows, dominated by workflows with 0–1 sampled runs in the last 30 days (unreliable Monte Carlo, several with sampled_runs: 0 and no monte-carlo object at all — e.g. Docs Explanatory Diagram Generator, Ambient Context, AW Failures, EU CRA Advisor sub-workflows, Enterprise canary/review-stress, SVG Visual Audit).
• Several workflows also show wide P90/P50 intervals (>2x), reflecting high variance in per-run AIC (e.g. Daily Dashboard Language Specification Review, Dependabot, Optimization, Advisory).
• No missing top-spend workflows detected — the 3 largest observed spenders (Multi-Device Docs Tester, CAO Dashboard Review, Daily Dashboard Language Renderer) all have reliable (is_reliable=true) Monte Carlo projections with ≥12 sampled runs.

📈 Trends

This is the first day a snapshot has been recorded in repo-memory for this repository (rolling-summary.json currently has 1 entry), so the historical trend chart is skipped — fewer than 2 data points are available. Trend comparisons will begin appearing once a second daily snapshot is collected.

💡 Observations

• Heavy hitter: Multi-Device Docs Tester accounts for 585.78 / 953.56 AIC (~61.4% of total spend) — well above the 30% threshold.
• High avg AIC/run (>1.00 AIC): Multi-Device Docs Tester (48.81), CAO Dashboard Review (20.49), Daily Dashboard Language Renderer (17.41).
• Daily Dashboard Language Renderer recorded 1 error across 7 runs; no other elevated error/warning rates observed.
• The 15 remaining workflows in this window (mostly dispatched sub-jobs of Dependabot, EU CRA Advisor, Ambient Context, Optimization, Advisory, AW Failures/Maintenance families) logged 0.0 AIC and 0 tokens in this snapshot — likely orchestrator/dispatch-only runs whose cost is attributed elsewhere, or lightweight no-op runs. Worth a closer look if this persists across future snapshots.

Data snapshot: memory/token-audit/githubnext__central-agentic-ops__2026-08-28.json

Control Plane

• Correlation ID: 33205281790-1
• Central repository: githubnext/central-agentic-ops
• Control plane run URL: https://github.com/githubnext/central-agentic-ops/actions/runs/33205281790
• Review mode: this issue is published to the safe-output repository (same as target repository in this run).

Generated by 🔍 Optimization / AI Credit Auditor · copilot · auto · 65.7 AIC · ⌖ 8.13 AIC · ⊞ 19.6K · ◷

gh aw add githubnext/central-agentic-ops/.github/workflows/optimization-ai-credit-auditor.md@main

• expires on Aug 31, 2026, 8:01 PM UTC