[copilot-token-audit] Daily Copilot Token Usage Audit — 2026-04-16

[github-actions[bot]]

📊 Executive Summary

• Period: Last 30 days (snapshot date: 2026-04-16)
• Total completed runs: 98 (out of 100 total, 2 in-progress excluded)
• Total tokens: 88,283,048
• Total cost: $0.00
• Total Actions minutes: 833 min
• Active workflows: 53

Day-over-day: +18.5% tokens vs yesterday (2026-04-15: 74,512,455 → 2026-04-16: 88,283,048), same run count (98).

---

🏆 Top 5 Workflows by Token Usage

Workflow	Runs	Total Tokens	Avg Tokens/Run	Total Cost	Errors
Test Quality Sentinel	30	11,503,281	383,442	$0.00	0
Daily Community Attribution Updater	1	7,461,657	7,461,657	$0.00	⚠️ 1
Contribution Check	4	6,063,734	1,515,933	$0.00	⚠️ 4
Daily Syntax Error Quality Check	1	5,128,866	5,128,866	$0.00	0
Copilot CLI Deep Research Agent	1	4,401,775	4,401,775	$0.00	0

---

📈 Trends

Token Usage by Workflow (Top 15)

Historical Daily Token Usage

---

Full Per-Workflow Breakdown (53 workflows)

Workflow	Runs	Total Tokens	Avg Tokens	Errors	Warnings
Test Quality Sentinel	30	11,503,281	383,442	0	0
Daily Community Attribution Updater	1	7,461,657	7,461,657	1	0
Contribution Check	4	6,063,734	1,515,933	4	0
Daily Syntax Error Quality Check	1	5,128,866	5,128,866	0	0
Copilot CLI Deep Research Agent	1	4,401,775	4,401,775	0	0
Agent Persona Explorer	1	3,458,396	3,458,396	0	0
jsweep - JavaScript Unbloater	1	3,086,630	3,086,630	0	0
Package Specification Librarian	1	3,057,769	3,057,769	0	0
Daily Compiler Quality Check	1	3,049,793	3,049,793	0	0
Daily DIFC Integrity-Filtered Events Analyzer	1	2,390,894	2,390,894	1	0
Workflow Normalizer	1	2,377,625	2,377,625	0	0
Code Simplifier	1	2,148,216	2,148,216	0	0
Copilot Token Usage Optimizer	1	1,844,652	1,844,652	0	0
Architecture Diagram Generator	1	1,714,313	1,714,313	0	0
Smoke Copilot	2	1,665,124	832,562	0	0
Q	1	1,621,770	1,621,770	0	0
Daily CLI Tools Exploratory Tester	1	1,611,376	1,611,376	0	0
Daily MCP Tool Concurrency Analysis	1	1,358,600	1,358,600	0	0
Agent Performance Analyzer - Meta-Orchestrator	1	1,316,723	1,316,723	0	0
Copilot PR Conversation NLP Analysis	1	1,286,589	1,286,589	1	0
Functional Pragmatist	1	1,214,755	1,214,755	0	0
Daily Copilot PR Merged Report	1	1,187,935	1,187,935	0	0
Glossary Maintainer	1	1,186,130	1,186,130	0	0
Sub-Issue Closer	1	1,099,273	1,099,273	0	0
Package Specification Extractor	1	1,094,687	1,094,687	0	0
Issue Monster	3	1,093,380	364,460	0	0
Breaking Change Checker	1	1,021,835	1,021,835	0	0
Copilot PR Prompt Pattern Analysis	1	983,640	983,640	0	0
Daily Project Performance Summary Generator (Using MCP Scripts)	1	953,975	953,975	1	0
Daily Malicious Code Scan Agent	1	949,537	949,537	0	0
Dev	1	922,609	922,609	0	0
Release	1	912,391	912,391	0	0
Metrics Collector - Infrastructure Agent	1	867,996	867,996	0	0
Auto-Triage Issues	9	836,512	92,945	4	0
Architecture Guardian	1	834,775	834,775	0	0
Daily CLI Performance Agent	1	821,947	821,947	0	0
PR Triage Agent	2	796,409	398,204	0	0
Daily Regulatory Report Generator	1	676,532	676,532	0	0
Refactoring Cadence	1	661,674	661,674	0	0
GPL Dependency Cleaner (gpclean)	1	540,480	540,480	0	0
Daily Semgrep Scan	1	523,928	523,928	0	0
Issue Triage Agent	1	463,091	463,091	0	0
Daily Workflow Updater	1	460,484	460,484	0	0
Daily Safe Output Integrator	1	430,253	430,253	0	0
Daily Secrets Analysis Agent	1	380,954	380,954	0	0
Daily Hippo Learn	1	317,200	317,200	0	0
Agent Container Smoke Test	2	294,364	147,182	0	0
Draft PR Cleanup	1	107,812	107,812	0	0
GitHub Remote MCP Authentication Test	1	100,707	100,707	1	0
Daily Issues Report Generator	1	0	0	1	0
Daily News	1	0	0	1	0
The Daily Repository Chronicle	1	0	0	0	0
Super Linter Report	1	0	0	1	0

---

💡 Observations

• Test Quality Sentinel is the top consumer at 13.0% of total tokens (11.5M), running 30 times with an avg of 383K tokens/run — the only workflow with high-frequency runs.
• Daily Community Attribution Updater has the highest avg tokens per run at 7,461,657 — a single run consuming 8.5% of the period's total. Combined with a 100% error rate, this warrants investigation.
• Contribution Check consumed 4 errors across 4 runs (100% error rate) — every run errored.
• 9 workflows recorded errors on every single run (100% error rate): Daily Community Attribution Updater, Contribution Check, Daily DIFC Integrity-Filtered Events Analyzer, Copilot PR Conversation NLP Analysis, Daily Project Performance Summary Generator, GitHub Remote MCP Authentication Test, Daily Issues Report Generator, Daily News, Super Linter Report.
• 48 out of 53 workflows (91%) had avg tokens per run exceeding 100,000 tokens.
• Token usage is up +18.5% vs yesterday despite the same number of completed runs, driven by a handful of high-consumption single-run workflows.

Data snapshot: memory/token-audit/2026-04-16.json

References:

• §24508228891 — This audit run

Generated by Daily Copilot Token Usage Audit · ● 1.3M · ◷

• expires on Apr 19, 2026, 12:25 PM UTC

[github-actions[bot]]

This discussion has been marked as outdated by Daily Copilot Token Usage Audit.

A newer discussion is available at Discussion #26856.