Skip to content

[agentic-token-audit] Daily AIC Usage Audit β€” 2026-06-09Β #38109

Description

@github-actions

πŸ“Š Executive Summary

  • Period: last 24 hours (2026-06-08 to 2026-06-09)
  • Total runs: 95 (completed) out of 100 total (5 in-progress)
  • Total AIC: 14,462.47
  • Total Actions minutes: 1,178.0 min
  • Active workflows: 53

πŸ† Top 5 Workflows by AIC Usage

Workflow Runs Total AIC Avg AIC
Matt Pocock Skills Reviewer 6 1,334.62 222.44
Test Quality Sentinel 6 1,255.05 209.18
Daily CLI Tools Exploratory Tester 1 923.26 923.26
Contribution Check 2 805.00 402.50
Package Specification Extractor 1 700.19 700.19

πŸ“ˆ Trends

AIC by Workflow

Historical AIC Trend

Historical AIC data covers 2 days (rolling summary migrated to AIC metric as of 2026-06-08):

  • 2026-06-08: 13,224.34 AIC Β· 97 runs
  • 2026-06-09: 14,462.47 AIC Β· 95 runs β†’ +9.4% AIC day-over-day despite 2 fewer runs, indicating higher per-run cost.
Full Per-Workflow Breakdown
Workflow Runs Total AIC Avg AIC Avg Turns Action Min Errors Warnings
Matt Pocock Skills Reviewer 6 1,334.62 222.44 14.7 69.0 0 0
Test Quality Sentinel 6 1,255.05 209.18 18.7 68.0 0 0
Daily CLI Tools Exploratory Tester 1 923.26 923.26 46.0 19.0 0 0
Contribution Check 2 805.00 402.50 37.5 25.0 0 0
Package Specification Extractor 1 700.19 700.19 35.0 12.0 0 0
Daily Agentic Workflow AIC Usage Audit 2 688.19 344.10 22.5 32.0 0 0
Impact Efficiency Report 1 660.48 660.48 36.0 19.0 0 0
PR Description Updater 8 627.71 78.46 10.1 63.0 0 0
Glossary Maintainer 1 624.72 624.72 27.0 13.0 0 0
Terminal Stylist 1 540.50 540.50 27.0 10.0 0 0
GitHub API Consumption Report Agent 1 453.73 453.73 53.0 42.0 0 0
Daily Semgrep Scan 1 412.00 412.00 37.0 11.0 0 0
PR Sous Chef 6 392.52 65.42 23.3 58.0 1 0
Daily Rendering Scripts Verifier 1 383.36 383.36 73.0 16.0 0 0
Workflow Health Manager - Meta-Orchestrator 1 349.82 349.82 27.0 11.0 0 0
Daily A/B Testing Advisor 1 327.98 327.98 29.0 12.0 0 0
Instructions Janitor 1 309.30 309.30 43.0 14.0 0 0
Daily OTel Instrumentation Advisor 1 295.34 295.34 19.0 12.0 0 0
Copilot Session Insights 1 287.51 287.51 41.0 36.0 0 0
Daily AW Cross-Repo Compile Check 1 275.00 275.00 46.0 76.0 0 0
Static Analysis Report 1 274.57 274.57 39.0 15.0 0 0
Daily AgentRx Trace Optimizer 1 264.66 264.66 34.0 14.0 0 0
Functional Pragmatist 1 225.89 225.89 14.0 9.0 0 0
Daily Documentation Updater 1 225.23 225.23 59.0 11.0 0 0
OTLP Data Quality Validator 1 204.60 204.60 13.0 9.0 0 0
[aw] Failure Investigator (6h) 1 204.21 204.21 30.0 15.0 0 0
Go Fan 1 189.28 189.28 37.0 14.0 0 0
Daily Hippo Learn 1 163.36 163.36 15.0 9.0 0 0
Design Decision Gate πŸ—οΈ 6 161.82 26.97 3.0 40.0 0 0
Copilot Agent Prompt Clustering Analysis 1 161.66 161.66 1.0 29.0 0 0
CLI Version Checker 1 123.07 123.07 17.0 7.0 0 0
Schema Consistency Checker 1 120.72 120.72 15.0 8.0 0 0
Sub-Issue Closer 1 111.30 111.30 14.0 7.0 0 0
Daily Windows Terminal Integration Builder 1 103.68 103.68 8.0 12.0 0 0
Multi-Device Docs Tester 1 76.32 76.32 15.0 8.0 0 0
Auto-Triage Issues 3 49.70 16.57 9.0 26.0 0 0
Daily Safe Outputs Conformance Checker 1 41.21 41.21 7.0 7.0 0 0
Daily Choice Type Test 1 28.69 28.69 3.0 5.0 0 0
Dependabot Campaign 1 28.18 28.18 3.0 5.0 0 0
Dev 1 25.04 25.04 0.0 8.0 0 0
Duplicate Code Detector 1 24.56 24.56 0.0 6.0 0 0
Daily Max AI Credits Test (Intentionally Fails) 1 8.46 8.46 1.0 5.0 1 0
Smoke CI 6 0.00 0.00 0.0 33.0 0 0
Draft PR Cleanup 1 0.00 0.00 1.0 7.0 0 0
PR Code Quality Reviewer 6 0.00 0.00 1.0 93.0 2 0
Copilot PR Conversation NLP Analysis 1 0.00 0.00 1.0 22.0 0 0
Daily MCP Tool Concurrency Analysis 1 0.00 0.00 1.0 49.0 1 0
Daily Syntax Error Quality Check 1 0.00 0.00 1.0 9.0 0 0
Daily Workflow Updater 2 0.00 0.00 1.0 32.0 1 0
PR Triage Agent 1 0.00 0.00 1.0 9.0 0 0
Daily News 1 0.00 0.00 0.0 8.0 1 0
daily-experiment-report 1 0.00 0.00 1.0 25.0 0 0
GitHub Remote MCP Authentication Test 1 0.00 0.00 0.0 4.0 1 0

πŸ’‘ Observations

  • No heavy hitter: No single workflow exceeds 30% of total AIC (threshold: 4,338.74). The top workflow, Matt Pocock Skills Reviewer, accounts for 9.2%.
  • High avg AIC per run: 41 of 53 workflows exceed 10.00 avg AIC/run β€” primarily deep analysis, documentation, and multi-step review workflows, which is expected. Notable single-run outliers: Daily CLI Tools Exploratory Tester (923.26), Package Specification Extractor (700.19), Impact Efficiency Report (660.48).
  • Long turn chains: Daily Rendering Scripts Verifier averaged 73 turns/run β€” highest in the window. Candidates for turn-budget review.
  • Error flags: PR Code Quality Reviewer logged 2 errors across 6 runs (33% error rate) with 0 AIC β€” likely failing before model invocation. Daily News, GitHub Remote MCP Authentication Test, and Daily MCP Tool Concurrency Analysis each had 1 error on their sole run.
  • Zero-AIC workflows: 12 workflows completed with 0 AIC β€” these are either non-LLM CI jobs (Smoke CI) or workflows failing before model calls. Reviewing errors on PR Code Quality Reviewer is recommended.
  • Day-over-day trend: AIC rose +9.4% while run count dropped by 2 β€” per-run cost is increasing. Only 2 days of AIC history exist in rolling summary (legacy entries use token counts); trend tracking will improve over time.

Data snapshot: memory/token-audit/2026-06-09.json

Generated by Daily Agentic Workflow AIC Usage Audit Β· 238.1 AIC Β· βŒ– 14.7 AIC Β· ⊞ 34.9K Β· β—·

  • expires on Jun 12, 2026, 4:41 AM UTC-08:00

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions