Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyzes task complexity, identifies required agents, groups them into optimal batches, and executes batches sequentially (agents within each batch run in parallel). Each batch gets a dedicated context window. Results are written to files and synthesized by an orchestrator. Achieves 60-77% token sav
.claude/skills/miosa-osa-batch/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 40% | 0% |
> Execute multi-agent tasks using intelligent batching for token efficiency.
/batch "<task>" [--max-agents <n>] [--batch-size <n>]Analyzes task complexity, identifies required agents, groups them into optimal batches, and executes batches sequentially (agents within each batch run in parallel). Each batch gets a dedicated context window. Results are written to files and synthesized by an orchestrator. Achieves 60-77% token savings vs naive parallel execution.
work/batch1-results.md.work/batch2-results.md.| Complexity | Batch Size | Example | |------------|------------|---------| | 1-3 | 1-2 agents | Fix typo, add logging | | 4-5 | 3 agents | Add API endpoint with tests | | 6-7 | 5 agents | Build feature with frontend/backend | | 8-10 | 8 agents | Full system redesign |
bash# Full-stack feature /batch "Build user authentication with React frontend, Go backend, and tests" # Performance work /batch "Optimize database queries and add caching layer" # Security audit /batch "Security assessment of payment processing system"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 32,519 | 28,650 | -12% | 1 | 1 | 0% | 5,870 | 4,862 | -17% | 0 | 0 | — |
case-02 | fail→fail | 30,637 | 4,805 | -84% | 1 | 1 | 0% | 4,754 | 790 | -83% | 0 | 0 | — |
case-03 | fail→pass | 54,752 | 31,850 | -42% | 1 | 1 | 0% | 6,765 | 7,277 | +8% | 0 | 0 | — |
case-04 | fail→pass | 6,049 | 9,715 | +61% | 1 | 1 | 0% | 984 | 1,226 | +25% | 0 | 0 | — |
case-05 | fail→fail | 17,770 | 5,599 | -68% | 1 | 1 | 0% | 3,127 | 766 | -76% | 0 | 0 | — |
case-06 | fail→fail | 44,659 | 38,044 | -15% | 1 | 1 | 0% | 7,787 | 7,074 | -9% | 0 | 0 | — |
case-07 | fail→pass | 23,392 | 34,074 | +46% | 1 | 1 | 0% | 3,351 | 5,275 | +57% | 0 | 0 | — |
case-08 | fail→pass | 15,148 | 20,154 | +33% | 1 | 1 | 0% | 2,128 | 2,979 | +40% | 0 | 0 | — |
case-09 | pass→pass | 27,915 | 26,270 | -6% | 1 | 1 | 0% | 4,205 | 5,408 | +29% | 0 | 0 | — |
case-10 | fail→fail | 32,095 | 41,743 | +30% | 1 | 1 | 0% | 5,604 | 7,109 | +27% | 0 | 0 | — |
case-11 | fail→pass | 22,883 | 20,128 | -12% | 1 | 1 | 0% | 4,166 | 4,356 | +5% | 0 | 0 | — |
case-12 | fail→fail | 122,021 | 47,768 | -61% | 1 | 1 | 0% | 1,224 | 1,178 | -4% | 0 | 0 | — |
case-13 | fail→fail | 24,172 | 6,988 | -71% | 1 | 1 | 0% | 4,009 | 734 | -82% | 0 | 0 | — |
case-14 | fail→pass | 31,291 | 42,448 | +36% | 1 | 1 | 0% | 8,203 | 5,244 | -36% | 0 | 0 | — |
case-15 | fail→fail | 46,462 | 24,021 | -48% | 1 | 1 | 0% | 8,211 | 5,682 | -31% | 0 | 0 | — |
case-16 | fail→pass | 21,748 | 43,400 | +100% | 1 | 1 | 0% | 3,537 | 6,383 | +80% | 0 | 0 | — |
case-17 | pass→pass | 28,808 | 25,524 | -11% | 1 | 1 | 0% | 4,278 | 5,731 | +34% | 0 | 0 | — |
case-18 | pass→fail | 14,723 | 5,340 | -64% | 1 | 1 | 0% | 3,155 | 719 | -77% | 0 | 0 | — |
case-19 | fail→pass | 36,323 | 48,512 | +34% | 1 | 1 | 0% | 8,216 | 8,604 | +5% | 0 | 0 | — |
case-20 | fail→pass | 40,375 | 22,610 | -44% | 1 | 1 | 0% | 5,844 | 3,568 | -39% | 0 | 0 | — |
case-21 | pass→pass | 13,743 | 17,574 | +28% | 1 | 1 | 0% | 2,587 | 3,038 | +17% | 0 | 0 | — |
case-22 | pass→pass | 18,051 | 35,491 | +97% | 1 | 1 | 0% | 3,316 | 3,617 | +9% | 0 | 0 | — |
case-23 | pass→pass | 19,906 | 15,359 | -23% | 1 | 1 | 0% | 3,047 | 3,609 | +18% | 0 | 0 | — |
case-24 | pass→pass | 21,473 | 14,672 | -32% | 1 | 1 | 0% | 3,204 | 3,427 | +7% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 20 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +38 percentage points is the difference between those two pass rates over the 20 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.