Overview
Period: 2026-05-15 13:31 UTC to 2026-05-16 01:12 UTC
Runs analyzed: 50 (35 had token data)
Total tokens: 7,435,753 across all workflows
Total effective tokens: 1,037,683 (cost-normalized, ~86% cache savings)
Estimated total cost: $3.65
Total Actions minutes: 151 min
Total turns: 226
Workflow Summary
| Workflow |
Runs |
Total Tokens |
Avg Tokens/Run |
Total Cost |
Avg Cost/Run |
Total Turns |
Avg Turns/Run |
| Smoke Claude |
27 |
5,589K |
207K |
$1.92 |
$0.07 |
177 |
6.6 |
| Security Guard |
23 |
1,847K |
80K |
$1.73 |
$0.08 |
49 |
2.1 |
🔍 Optimization Opportunities
-
Security Guard — 15 of 23 runs (65%) have no token_usage data
- These successful runs show
token_usage: null, suggesting the api-proxy is not consistently active or token reporting is not wired up for all execution paths
- Recommendation: Verify that all Security Guard workflow runs use
--enable-api-proxy to enable token tracking
-
Smoke Claude — 3/27 runs reported errors (11% error rate)
-
Security Guard — 1/23 runs had an error
- Error run: 25921181857 (coincides with Smoke Claude error in same batch)
- Likely a shared infrastructure transient failure
Per-Workflow Details
Smoke Claude
- Runs: 27 (recent, 25947365843, 25946257244 and 24 more)
- Total tokens: 5,589,168 (avg 207,006/run)
- Effective tokens: 542,852 (avg 20,105/run) — 90.3% reduction via caching
- Estimated cost: $1.92 (avg $0.07/run)
- Turns: 177 total (avg 6.6/run)
- Model:
claude-haiku-4-5-20251001 (Anthropic)
- Cache analysis: 931K write tokens, 4,613K read tokens — avg efficiency 98.2% ✅ (excellent reuse)
- Error rate: 3/27 runs (11%)
Security Guard
- Runs: 23 (25946257216, 25946132915, 25932869446 and 20 more)
- Total tokens: 1,846,585 (avg 80,286/run, from runs with token data)
- Effective tokens: 494,831 (avg 21,514/run)
- Estimated cost: $1.73 (avg $0.08/run)
- Turns: 49 total (avg 2.1/run)
- Models:
claude-haiku-4-5-20251001 + claude-sonnet-4-6 (Anthropic)
- Cache analysis: 273K write tokens, 1,559K read tokens — avg efficiency 97.4% ✅ (excellent reuse)
- Error rate: 1/23 runs (4%)
Workflows Without Token Data
15 Security Guard runs had token_usage: null. All completed successfully, suggesting the api-proxy may not be consistently enabled or token reporting is not active for all invocations:
Historical Trend
This is the first Claude token usage report for this period. No prior reports found to compare against.
References: Workflow Run §25957645512
Generated by Daily Claude Token Usage Analyzer · ● 4.1M · ◷
Overview
Period: 2026-05-15 13:31 UTC to 2026-05-16 01:12 UTC
Runs analyzed: 50 (35 had token data)
Total tokens: 7,435,753 across all workflows
Total effective tokens: 1,037,683 (cost-normalized, ~86% cache savings)
Estimated total cost: $3.65
Total Actions minutes: 151 min
Total turns: 226
Workflow Summary
🔍 Optimization Opportunities
Security Guard — 15 of 23 runs (65%) have no
token_usagedatatoken_usage: null, suggesting the api-proxy is not consistently active or token reporting is not wired up for all execution paths--enable-api-proxyto enable token trackingSmoke Claude — 3/27 runs reported errors (11% error rate)
Security Guard — 1/23 runs had an error
Per-Workflow Details
Smoke Claude
claude-haiku-4-5-20251001(Anthropic)Security Guard
claude-haiku-4-5-20251001+claude-sonnet-4-6(Anthropic)Workflows Without Token Data
15 Security Guard runs had
token_usage: null. All completed successfully, suggesting the api-proxy may not be consistently enabled or token reporting is not active for all invocations:Historical Trend
This is the first Claude token usage report for this period. No prior reports found to compare against.
References: Workflow Run §25957645512