Overview
Period: 2026-05-30 to 2026-05-31
Runs analyzed: 41 (29 had token data)
Total tokens: 9,388K across all workflows
Estimated total cost: $4.24
Total Actions minutes: 97 min
Workflow Summary
| Workflow |
Runs |
Runs w/ Tokens |
Total Tokens |
Avg Tokens |
Total Cost |
Avg Cost |
Total Turns |
| Smoke Claude |
22 |
22 |
7,400K |
336K |
$2.11 |
$0.10 |
120 |
| Security Guard |
18 |
6 |
1,882K |
314K |
$2.05 |
$0.34 |
29 |
| Documentation Maintainer |
1 |
1 |
106K |
106K |
$0.08 |
$0.08 |
2 |
🔍 Optimization Opportunities
-
Security Guard — $0.34/run (for runs with token data), 12/18 runs missing token data
- 12 of 18 runs report 0 tokens — these may be conditional/skipped agent executions or runs where api-proxy was not active
- For the 6 runs with token data, avg is 313K tokens at $0.34/run — high cost per run
- Investigate whether the high-turn runs (7–8 turns) can be reduced with more targeted prompting
-
Smoke Claude — $0.10/run average, 336K avg tokens/run
- Consistent 4–7 turns per run with predictable token usage (245K–433K range)
- Running at high volume (22 runs in 2 days) — consider if all runs are necessary or if frequency can be reduced
- Token usage is stable, suggesting good prompt hygiene; no immediate optimization needed
-
No prompt cache data available — token_usage_summary was not populated for any workflow
- Prompt caching (Anthropic cache read/write tokens) could significantly reduce costs on repetitive workflows like Smoke Claude
- Enable detailed token tracking via api-proxy to capture cache hit rates
Per-Workflow Details
Smoke Claude
Security Guard
- Runs: 18 — all successful (12 with 0 tokens, 6 with token data)
- Total tokens: 1,882K (avg 314K/run for runs with data)
- Estimated cost: $2.05 (avg $0.34/run for runs with data)
- Turns: 29 total (avg 4.8/run for runs with data; range 2–8)
- Cache analysis: No token_usage_summary data available
- Error rate: 0/18 runs
- Run links (with token data): 26687379782 · 26687026287 · 26685872057 · 26685602930 · 26685273422 · 26685102617
Documentation Maintainer
- Runs: 1 — successful
- Total tokens: 106K
- Estimated cost: $0.08
- Turns: 2
- Cache analysis: No token_usage_summary data available
- Error rate: 0/1 runs
- Run link: 26704883749
Workflows Without Token Data
12 Security Guard runs reported 0 tokens. These may be conditional/skipped agent executions (e.g., no triggering PR changes), or runs where the api-proxy was not active:
Historical Trend
This is the first Claude token usage report (no previous reports found with the claude-token-usage-report label).
References: §26708637938
Generated by Daily Claude Token Usage Analyzer · sonnet46 968.2K · ◷
Overview
Period: 2026-05-30 to 2026-05-31
Runs analyzed: 41 (29 had token data)
Total tokens: 9,388K across all workflows
Estimated total cost: $4.24
Total Actions minutes: 97 min
Workflow Summary
🔍 Optimization Opportunities
Security Guard — $0.34/run (for runs with token data), 12/18 runs missing token data
Smoke Claude — $0.10/run average, 336K avg tokens/run
No prompt cache data available —
token_usage_summarywas not populated for any workflowPer-Workflow Details
Smoke Claude
Security Guard
Documentation Maintainer
Workflows Without Token Data
12 Security Guard runs reported 0 tokens. These may be conditional/skipped agent executions (e.g., no triggering PR changes), or runs where the api-proxy was not active:
Historical Trend
This is the first Claude token usage report (no previous reports found with the
claude-token-usage-reportlabel).References: §26708637938