Overview
Period: 2026-06-02 to 2026-06-03
Runs analyzed: 50 (50 had token data)
Total tokens: 18,794K across all workflows
Estimated total cost: $12.91
Total Actions minutes: 209 min
⚠️ Cost increased +47% vs previous report ($8.80 → $12.91) and tokens increased +79% (10,527K → 18,794K) — driven primarily by Security Guard (48% failure rate consuming tokens without useful output).
Workflow Summary
| Workflow |
Runs |
Total Tokens |
Avg Tokens |
Cost |
Avg Cost |
Turns |
| Security Guard |
25 |
12,650K |
506K |
$10.70 |
$0.43 |
189 |
| Smoke Claude |
25 |
6,144K |
246K |
$2.21 |
$0.09 |
101 |
🔍 Optimization Opportunities
-
Security Guard — $0.43/run average, 48% failure rate (12/25 runs failed)
- 12 runs encountered errors, consuming significant tokens on failed work
- High avg turns (7.6/run) combined with failures suggests the agent is struggling and retrying
- Investigate root cause of failures to reduce wasted token spend
- Consider adding pre-checks or guard conditions to fail fast before consuming many turns
-
Security Guard — 506K avg tokens/run (threshold: 100K)
- Token count is 5× above the high-usage threshold
- Review system prompt size and MCP tool surface area
- Consider pre-computing deterministic checks (e.g., static analysis) outside the LLM
-
Security Guard — High error rate (48% failure rate)
- Failures are wasting an estimated ~$4+ in tokens per day
- Short-circuiting on error conditions earlier could cut costs significantly
Per-Workflow Details
Security Guard
- Runs: 25
- Total tokens: 12,650K (avg 506K/run)
- Estimated cost: $10.70 (avg $0.43/run)
- Turns: 189 total (avg 7.6/run)
- Cache analysis: No cache token data available (
token_usage_summary not present)
- Error rate: 12/25 runs failed (48%)
- Conclusions: 13 success, 12 failure
- Recent runs: §26868586783 · §26867322267 · §26864943353
Smoke Claude
- Runs: 25
- Total tokens: 6,144K (avg 246K/run)
- Estimated cost: $2.21 (avg $0.09/run)
- Turns: 101 total (avg 4.0/run)
- Cache analysis: No cache token data available (
token_usage_summary not present)
- Error rate: 0/25 runs (100% success)
- Recent runs: §26868586774 · §26867322233 · §26864943342
Workflows Without Token Data
All 50 runs had token usage data present. No workflows are missing instrumentation.
Historical Trend
| Date |
Total Tokens |
Total Cost |
Runs |
| 2026-06-03 |
18,794K |
$12.91 |
50 |
| 2026-06-02 |
10,527K |
$8.80 |
50 |
↑ +47% cost, +79% tokens vs yesterday — Security Guard failure rate is the primary driver of increased spend.
Previous Report
📊 Claude Token Usage Report 2026-06-02
References:
Generated by Daily Claude Token Usage Analyzer · sonnet46 920.1K · ◷
Overview
Period: 2026-06-02 to 2026-06-03
Runs analyzed: 50 (50 had token data)
Total tokens: 18,794K across all workflows
Estimated total cost: $12.91
Total Actions minutes: 209 min
Workflow Summary
🔍 Optimization Opportunities
Security Guard — $0.43/run average, 48% failure rate (12/25 runs failed)
Security Guard — 506K avg tokens/run (threshold: 100K)
Security Guard — High error rate (48% failure rate)
Per-Workflow Details
Security Guard
token_usage_summarynot present)Smoke Claude
token_usage_summarynot present)Workflows Without Token Data
All 50 runs had token usage data present. No workflows are missing instrumentation.
Historical Trend
↑ +47% cost, +79% tokens vs yesterday — Security Guard failure rate is the primary driver of increased spend.
Previous Report
📊 Claude Token Usage Report 2026-06-02
References: