Overview
Period: 2026-08-28 to 2026-08-29 (last 24h)
Runs analyzed: 5 (5 had token data, 0 had cost/turn data)
Total tokens: 25K across all workflows
Estimated total cost: $0.00 (not available — see note below)
Total 操作 minutes: 29 min
Workflow Summary
| Workflow |
Runs |
Total Tokens |
Avg Tokens |
Cost |
Avg Cost |
Turns |
| Smoke Claude |
5 |
25K |
5.1K |
N/A |
N/A |
N/A |
🔍 Optimization Opportunities
No optimization flags triggered. Token usage per run (~5K tokens) is well below the 100K threshold, and all 5 runs completed successfully with 0 errors/warnings. Action minutes (4–8 min/run) are modest.
Data gap: estimated_cost, turns, and token_usage_summary (model/cache breakdown) are null for all runs, meaning the api-proxy sidecar's token-usage instrumentation isn't populating cost/cache data for this workflow. Without cache read/write figures, we can't assess Anthropic prompt-caching efficiency. Recommend verifying --enable-api-proxy is active for smoke-claude.lock.yml and that its cost-tracking fields are being recorded.
Per-Workflow Details
Smoke Claude
- Runs: 5 — §33225103470, §33188263454, §33188200527, §33188145850, §33187887609
- Total tokens: 25,276 (avg 5,055/run)
- Estimated cost: not available (null in all runs)
- Turns: not available (null in all runs)
- Model breakdown: not available (
token_usage_summary null in all runs)
- Cache analysis: not available
- Error rate: 0/5 runs
- Action minutes: 29 total (avg 5.8/run)
Workflows Without Token Data
All 5 runs have non-null token_usage, so none are fully missing token data. However, all 5 are missing estimated_cost, turns, and token_usage_summary — these fields need api-proxy cost-instrumentation to populate for the "Smoke Claude" workflow.
Historical Trend
No previous Claude token usage report issues were found (searched issues labeled claude-token-usage-report). This is the first Claude token usage report.
References:
Generated by Daily Claude Token Usage Analyzer · copilot · auto · 20.5 AIC · ⊞ 13.4K · ◷
Overview
Period: 2026-08-28 to 2026-08-29 (last 24h)
Runs analyzed: 5 (5 had token data, 0 had cost/turn data)
Total tokens: 25K across all workflows
Estimated total cost: $0.00 (not available — see note below)
Total 操作 minutes: 29 min
Workflow Summary
🔍 Optimization Opportunities
No optimization flags triggered. Token usage per run (~5K tokens) is well below the 100K threshold, and all 5 runs completed successfully with 0 errors/warnings. Action minutes (4–8 min/run) are modest.
Data gap:
estimated_cost,turns, andtoken_usage_summary(model/cache breakdown) arenullfor all runs, meaning the api-proxy sidecar's token-usage instrumentation isn't populating cost/cache data for this workflow. Without cache read/write figures, we can't assess Anthropic prompt-caching efficiency. Recommend verifying--enable-api-proxyis active forsmoke-claude.lock.ymland that its cost-tracking fields are being recorded.Per-Workflow Details
Smoke Claude
token_usage_summarynull in all runs)Workflows Without Token Data
All 5 runs have non-null
token_usage, so none are fully missing token data. However, all 5 are missingestimated_cost,turns, andtoken_usage_summary— these fields need api-proxy cost-instrumentation to populate for the "Smoke Claude" workflow.Historical Trend
No previous Claude token usage report issues were found (searched issues labeled
claude-token-usage-report). This is the first Claude token usage report.References: