Overview
Period: 2026-08-21 to 2026-08-22
Runs analyzed: 6 (6 had token data)
Total tokens: 27K across all workflows
Estimated total cost: not available (no estimated_cost / token_usage_summary data in this dataset)
Total Actions minutes: 28 min
Only one Claude-engine workflow ran in this period: Smoke Claude. All 6 runs completed successfully with no errors or warnings.
Workflow Summary
| Workflow |
Runs |
Total Tokens |
Avg Tokens |
Cost |
Avg Cost |
Turns |
| Smoke Claude |
6 |
27K |
4.6K |
n/a |
n/a |
n/a |
🔍 Optimization Opportunities
No optimization flags were triggered:
- Avg tokens/run (~4.6K) is well below the 100K threshold.
- No cost data was available to evaluate the >$1.00/run threshold.
- Turn counts were not populated in this dataset, so the >15 turns/run check could not be evaluated.
- Error/warning counts were 0 across all runs (0% error rate).
- No cache read/write breakdown (
token_usage_summary) was present for any run, so prompt-caching efficiency cannot be assessed for this workflow.
Per-Workflow Details
Smoke Claude
- Runs: 6
- §32541361056 — 4,495 tokens, 4.0m
- §32532650018 — 4,589 tokens, 4.0m
- §32505871059 — 4,640 tokens, 3.7m
- §32505780977 — 4,614 tokens, 3.8m
- §32502435552 — 4,614 tokens, 4.2m
- §32482658496 — 4,455 tokens, 4.3m
- Total tokens: 27,407 (avg 4,568/run)
- Estimated cost: not available — no
estimated_cost field populated in the log data
- Turns: not available
- Model breakdown: not available (
token_usage_summary was null for all runs)
- Cache analysis: not available
- Error rate: 0/6 runs
Workflows Without Token Data
None — all 6 runs for the only Claude workflow observed (Smoke Claude) had non-zero token_usage. However, note that estimated_cost, turns, error_count, warning_count, and token_usage_summary were all null/absent across every run, suggesting the api-proxy sidecar's cost/cache instrumentation is either not enabled for this workflow or not being captured in the gh aw logs --json export for this run set.
Historical Trend
Previous Claude token usage report issues exist in this repository (labeled claude-token-usage-report), but their content could not be read in this session due to an integrity-policy restriction on those specific issues. As a result, a direct trend comparison against prior reports isn't possible this run. Token usage across the 6 runs analyzed here was stable, ranging narrowly from 4,455 to 4,640 tokens per run with no apparent drift.
Previous Report
Unable to retrieve/link the most recent previous report due to an integrity-policy restriction preventing this session from reading those issues.
References:
Generated by Daily Claude Token Usage Analyzer · auto · 27.5 AIC · ⊞ 13.4K · ◷
Overview
Period: 2026-08-21 to 2026-08-22
Runs analyzed: 6 (6 had token data)
Total tokens: 27K across all workflows
Estimated total cost: not available (no
estimated_cost/token_usage_summarydata in this dataset)Total Actions minutes: 28 min
Only one Claude-engine workflow ran in this period: Smoke Claude. All 6 runs completed successfully with no errors or warnings.
Workflow Summary
🔍 Optimization Opportunities
No optimization flags were triggered:
token_usage_summary) was present for any run, so prompt-caching efficiency cannot be assessed for this workflow.Per-Workflow Details
Smoke Claude
estimated_costfield populated in the log datatoken_usage_summarywas null for all runs)Workflows Without Token Data
None — all 6 runs for the only Claude workflow observed (Smoke Claude) had non-zero
token_usage. However, note thatestimated_cost,turns,error_count,warning_count, andtoken_usage_summarywere allnull/absent across every run, suggesting the api-proxy sidecar's cost/cache instrumentation is either not enabled for this workflow or not being captured in thegh aw logs --jsonexport for this run set.Historical Trend
Previous Claude token usage report issues exist in this repository (labeled
claude-token-usage-report), but their content could not be read in this session due to an integrity-policy restriction on those specific issues. As a result, a direct trend comparison against prior reports isn't possible this run. Token usage across the 6 runs analyzed here was stable, ranging narrowly from 4,455 to 4,640 tokens per run with no apparent drift.Previous Report
Unable to retrieve/link the most recent previous report due to an integrity-policy restriction preventing this session from reading those issues.
References: