📊 Copilot Token Usage Report2026-04-20

[github-actions]

Overview

Period: 2026-04-19 to 2026-04-20
Runs analyzed: 30 (29 had token data)
Total tokens: 8,949,198 raw · 10,776,131 effective (cost-normalized)
Estimated total cost: $0.00 (api-proxy cost tracking not yet emitting estimated_cost)
Total Actions minutes: 222 min
Total errors: 4 · Warnings: 0

💡 All token data sourced from token_usage_summary firewall telemetry. estimated_cost is $0 for all runs — cost calculation may not yet be wired to the api-proxy output. Effective tokens (cache-adjusted) are used for ranking.

---

Workflow Summary

Sorted by total effective tokens (highest cost proxy available):

Workflow	Runs	Total Tokens	Avg Tokens/Run	Eff Tokens	Turns	Errors
Weekly Test Coverage Improver	1	1,384K	1,384K	1,562K	1	0
Build Test Suite	2	1,237K	618K	1,386K	2	0
Firewall Issue Dispatcher	4	1,023K	256K	1,140K	4	0
Smoke Services	4	1,025K	256K	1,128K	4	2
Documentation Maintainer	1	799K	799K	890K	1	0
CLI Flag Consistency Checker	1	667K*	667K*	763K	1	1
Smoke Copilot BYOK	4	595K	149K	657K	4	0
Issue Duplication Detector	2	487K	244K	539K	2	0
Smoke Copilot	4	487K	122K	537K	4	1
Pelis Agent Factory Advisor	1	450K	450K	507K	1	0
Daily Security Review and Threat Modeling	1	419K	419K	475K	1	0
Daily Claude Token Optimization Advisor	1	417K	417K	479K	1	0
CI/CD Pipelines and Integration Tests Gap Assessment	1	340K	340K	385K	1	0
Daily Claude Token Usage Analyzer	1	168K	168K	197K	1	0
Dependency Security Monitor	1	119K	119K	132K	1	0
Daily Copilot Token Usage Analyzer	1	0	0	0	0	0

*CLI Flag Consistency Checker: token_usage field was null; tokens derived from token_usage_summary (input: 654K + output: 13K).

---

🔍 Optimization Opportunities

1. Weekly Test Coverage Improver — 1,384K tokens/run, 29 requests, 227s LLM time

   • Highest single-run token count. With 1,296K cache-read tokens (94% cache hit rate), caching is already working well.
   • Review if all 29 LLM requests are necessary or if some can be pre-computed deterministically.

2. Build Test Suite — avg 618K tokens/run (2 runs)

   • Large per-run average. 731K and 319K cache-read tokens across the two runs — cache efficiency varies.
   • Consider whether test analysis prompts can be made more concise.

3. Documentation Maintainer — 799K tokens in a single run, 22 requests

   • High request count (22) for a single run suggests iterative analysis. 745K cache-read tokens (93% cache rate) — very efficient.
   • Low output (5K output tokens vs 794K input) suggests mostly reading. Consider reducing context window fed per request.

4. Smoke Services — 2 errors in 4 runs (50% error rate) ⚠️

   • Highest error rate of any workflow. Worth investigating reliability issues.

5. CLI Flag Consistency Checker — 1 error in 1 run, token_usage field null

   • Check if api-proxy is correctly populating token_usage in structured output.

---

Per-Workflow Details

Weekly Test Coverage Improver

• Runs: 1 — §24651011761
• Total tokens: 1,384K (input: 1,368K, output: 16K)
• Cache-read tokens: 1,296K (94% cache efficiency)
• Effective tokens: 1,562K
• Requests: 29 · LLM time: 228s
• Model: claude-sonnet-4.6 (Copilot)
• Errors: 0

Build Test Suite

• Runs: 2 — §24629144022, §24640001477
• Total tokens: 1,237K (avg 618K/run)
• Cache-read tokens: 1,051K total
• Effective tokens: 1,386K
• Requests: 23 total
• Errors: 0

Firewall Issue Dispatcher

• Runs: 4 — §24629043408, §24636036745, §24643657777, §24653062778
• Total tokens: 1,023K (avg 256K/run)
• Effective tokens: 1,140K
• Turns: 4 (1 turn/run)
• Errors: 0

Smoke Services

• Runs: 4 — §24629144051, §24629439927, §24640001490, §24643750558
• Total tokens: 1,025K (avg 256K/run)
• Effective tokens: 1,128K
• Errors: 2 ⚠️

Documentation Maintainer

• Runs: 1 — §24629040615
• Total tokens: 799K (input: 794K, output: 5K)
• Cache-read tokens: 745K (94% cache rate)
• Effective tokens: 890K
• Requests: 22

CLI Flag Consistency Checker

• Runs: 1 — §24640133627
• token_usage: null (data gap) · Effective tokens from summary: 763K
• Input tokens: 654K · Output: 13K · Cache-read: 588K
• Requests: 17 · LLM time: 156s
• Errors: 1 ⚠️

Smoke Copilot BYOK

• Runs: 4 — §24629144035, §24629576215, §24640001470, §24643780249
• Total tokens: 595K (avg 149K/run)
• Effective tokens: 657K
• Errors: 0

Issue Duplication Detector

• Runs: 2 — §24636101824, §24653238165
• Total tokens: 487K (avg 244K/run)
• Effective tokens: 539K
• Errors: 0

Smoke Copilot

• Runs: 4 — §24629015495, §24629144049, §24640001488, §24643618644
• Total tokens: 487K (avg 122K/run)
• Effective tokens: 537K
• Errors: 1

Pelis Agent Factory Advisor

• Runs: 1 — §24639727444
• Total tokens: 450K (input: 443K, output: 7K)
• Cache-read tokens: 370K (83% cache rate)
• Effective tokens: 507K

Daily Security Review and Threat Modeling

• Runs: 1 — §24629763585
• Total tokens: 419K (input: 412K, output: 7K)
• Cache-read tokens: 339K (82% cache rate)
• Effective tokens: 475K

Daily Claude Token Optimization Advisor

• Runs: 1 — §24638122342
• Total tokens: 417K (input: 408K, output: 9K)
• Cache-read tokens: 355K (87% cache rate)
• Effective tokens: 479K

CI/CD Pipelines and Integration Tests Gap Assessment

• Runs: 1 — §24629384878
• Total tokens: 340K (input: 334K, output: 5K)
• Cache-read tokens: 289K (87% cache rate)
• Effective tokens: 385K

Daily Claude Token Usage Analyzer

• Runs: 1 — §24637808551
• Total tokens: 168K
• Effective tokens: 197K

Dependency Security Monitor

• Runs: 1 — §24641787623
• Total tokens: 119K
• Effective tokens: 132K

Cache Efficiency Summary

Cache hit rates (cache-read / total-input) across workflows with token data:

Workflow	Cache-Read Tokens	Input Tokens	Cache Rate
Weekly Test Coverage Improver	1,296K	1,368K	94% ✅
Documentation Maintainer	745K	794K	94% ✅
Firewall Issue Dispatcher (latest)	459K	500K	92% ✅
Daily Claude Token Optimization Advisor	355K	408K	87% ✅
CI/CD Gap Assessment	289K	334K	87% ✅
Build Test Suite (run 1)	732K	856K	86% ✅
Smoke Copilot BYOK	varies	—	~85%
CLI Flag Consistency Checker	588K	654K	90% ✅
Pelis Agent Factory Advisor	370K	443K	83% ✅
Daily Security Review	339K	412K	82% ✅

Overall cache efficiency is excellent across the board — system prompts and context are being reused effectively via prompt caching.

Workflows Without Token Data

• Daily Copilot Token Usage Analyzer (this run) — token_usage: 0, no token_usage_summary. Expected for current run.
• CLI Flag Consistency Checker — token_usage field was null. Token data available in token_usage_summary (663K tokens), so the firewall is tracking correctly; the top-level field aggregation may have a bug.

---

Historical Trend

This is the first Copilot token usage report — no prior reports found for comparison. Future reports will track week-over-week changes per workflow.

---

References:

• §24662200109 — This analyzer run
• §24651011761 — Weekly Test Coverage Improver (highest token run)
• §24629144022 — Build Test Suite

Generated by Daily Copilot Token Usage Analyzer · ● 297.1K · ◷