Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Manages parent/child agent relationships with task delegation and result aggregation. Supports sequential chains, parallel fans, conditional routing, retry logic, timeout handling, and YAML-based visual workflow definition.
.claude/skills/onewave-ai-sub-agent-orchestrator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-03 | ✓→✗ | ▼ Worse | -68% | 0% |
| case-12 | ✓→✓ | ＝ Same ✓ | 92% | 0% |
Design and execute multi-agent pipelines where each step is a different agent that depends on the previous one. Define roles, dependencies, and handoffs in YAML, then run sequential, parallel, conditional, loop, and map-reduce workflows with retry, timeout, and validation.
Unlike Agent Army (homogeneous parallel code changes) and Agent Swarm (homogeneous parallel data processing), this orchestrator coordinates heterogeneous pipelines where the output of A feeds the input of B.
references/patterns.md -- The six workflow patterns and the comparison to Agent Army/Swarm.references/workflow-schema.md -- Full YAML workflow definition language.references/examples.md -- Complete worked workflows (research-to-proposal, lead scoring).references/execution-engine.md -- Per-step execution model, retry, timeout, validation, edge cases.references/templates.md -- Reusable workflow scaffolds.references/visual-and-reporting.md -- Text diagrams and the execution report template.references/workflow-schema.md), then show it for approval.references/visual-and-reporting.md).references/execution-engine.md.references/execution-engine.md.references/visual-and-reporting.md. For partial or failed runs, report what completed, what failed, and any collected partial output.Match the task shape to a pattern, then scaffold from references/templates.md:
See references/patterns.md for diagrams and examples of each.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | pass→pass | 6,297 | 6,944 | +10% | 1 | 1 | 0% | 949 | 1,820 | +92% | 0 | 0 | — |
case-01 | fail→fail | 21,117 | 7,819 | -63% | 1 | 1 | 0% | 3,099 | 987 | -68% | 0 | 0 | — |
case-02 | fail→fail | 14,360 | 21,976 | +53% | 1 | 1 | 0% | 2,261 | 4,718 | +109% | 0 | 0 | — |
case-03 | pass→fail | 47,930 | 6,214 | -87% | 1 | 1 | 0% | 3,124 | 1,013 | -68% | 0 | 0 | — |
case-04 | pass→pass | 15,400 | 10,324 | -33% | 1 | 1 | 0% | 2,214 | 2,007 | -9% | 0 | 0 | — |
case-05 | pass→pass | 13,797 | 7,914 | -43% | 1 | 1 | 0% | 2,158 | 1,956 | -9% | 0 | 0 | — |
case-06 | pass→pass | 14,673 | 13,481 | -8% | 1 | 1 | 0% | 2,708 | 2,582 | -5% | 0 | 0 | — |
case-07 | pass→pass | 8,142 | 6,575 | -19% | 1 | 1 | 0% | 1,346 | 1,863 | +38% | 0 | 0 | — |
case-08 | pass→pass | 8,363 | 6,730 | -20% | 1 | 1 | 0% | 1,210 | 1,593 | +32% | 0 | 0 | — |
case-09 | pass→pass | 5,696 | 7,257 | +27% | 1 | 1 | 0% | 822 | 1,716 | +109% | 0 | 0 | — |
case-10 | pass→pass | 5,027 | 6,066 | +21% | 1 | 1 | 0% | 755 | 1,616 | +114% | 0 | 0 | — |
case-11 | pass→pass | 7,393 | 5,766 | -22% | 1 | 1 | 0% | 987 | 1,594 | +61% | 0 | 0 | — |
case-13 | pass→pass | 4,024 | 3,596 | -11% | 1 | 1 | 0% | 635 | 1,343 | +111% | 0 | 0 | — |
case-14 | pass→pass | 6,955 | 6,762 | -3% | 1 | 1 | 0% | 1,023 | 1,806 | +77% | 0 | 0 | — |
case-15 | pass→pass | 15,297 | 10,513 | -31% | 1 | 1 | 0% | 2,669 | 2,423 | -9% | 0 | 0 | — |
case-16 | fail→pass | 8,574 | 4,753 | -45% | 1 | 1 | 0% | 1,421 | 1,363 | -4% | 0 | 0 | — |
case-17 | pass→pass | 11,042 | 8,393 | -24% | 1 | 1 | 0% | 1,759 | 2,046 | +16% | 0 | 0 | — |
case-18 | pass→pass | 12,980 | 6,985 | -46% | 1 | 1 | 0% | 2,221 | 1,768 | -20% | 0 | 0 | — |
case-19 | pass→pass | 7,680 | 5,673 | -26% | 1 | 1 | 0% | 1,212 | 1,272 | +5% | 0 | 0 | — |
case-20 | fail→pass | 6,456 | 2,831 | -56% | 1 | 1 | 0% | 881 | 1,121 | +27% | 0 | 0 | — |
case-21 | pass→pass | 7,112 | 2,855 | -60% | 1 | 1 | 0% | 1,104 | 1,102 | -0% | 0 | 0 | — |
case-22 | fail→pass | 6,010 | 3,666 | -39% | 1 | 1 | 0% | 1,058 | 1,242 | +17% | 0 | 0 | — |
case-23 | pass→pass | 12,719 | 10,979 | -14% | 1 | 1 | 0% | 1,899 | 1,778 | -6% | 0 | 0 | — |
case-24 | pass→pass | 7,508 | 3,298 | -56% | 1 | 1 | 0% | 1,094 | 1,095 | +0% | 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 22 counted toward the lift figure. The other 2 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 +8 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are 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.