Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Creates an operation definition -- a reusable workflow that combines multiple agents and skills to achieve a specific outcome. Operations are the packaged units of work in the Workspace Protocol.
.claude/skills/miosa-osa-create-operation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -87% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -74% | 0% |
> Define a new operation (a packaged workflow with agents, skills, and deliverables).
/create-operation "<name>" --goal "<goal>" [--agents "<agent-list>"]Creates an operation definition -- a reusable workflow that combines multiple agents and skills to achieve a specific outcome. Operations are the packaged units of work in the Workspace Protocol.
operations/<name>/OPERATION.md.bash# Create a feature delivery operation /create-operation "feature-delivery" --goal "Ship a feature from spec to production" --agents "architect, backend-go, frontend-react, test-automator, devops-engineer" # Create a knowledge maintenance operation /create-operation "weekly-review" --goal "Friday review cycle" --agents "orchestrator" # Create a client onboarding operation /create-operation "client-onboard" --goal "Onboard new client from signup to first value"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | pass→fail | 13,447 | 16,409 | +22% | 1 | 1 | 0% | 2,329 | 3,069 | +32% | 0 | 0 | — |
case-22 | pass→pass | 14,428 | 11,510 | -20% | 1 | 1 | 0% | 2,898 | 2,449 | -15% | 0 | 0 | — |
case-01 | fail→fail | 13,040 | 14,087 | +8% | 1 | 1 | 0% | 2,472 | 1,348 | -45% | 0 | 0 | — |
case-02 | fail→fail | 13,706 | 8,889 | -35% | 1 | 1 | 0% | 2,463 | 1,881 | -24% | 0 | 0 | — |
case-03 | fail→fail | 14,563 | 8,943 | -39% | 1 | 1 | 0% | 2,717 | 1,920 | -29% | 0 | 0 | — |
case-04 | fail→fail | 21,031 | 3,315 | -84% | 1 | 1 | 0% | 3,640 | 889 | -76% | 0 | 0 | — |
case-05 | fail→pass | 12,379 | 9,308 | -25% | 1 | 1 | 0% | 2,372 | 1,987 | -16% | 0 | 0 | — |
case-06 | fail→pass | 27,479 | 1,994 | -93% | 1 | 1 | 0% | 4,522 | 592 | -87% | 0 | 0 | — |
case-07 | pass→pass | 9,591 | 2,543 | -73% | 1 | 1 | 0% | 1,373 | 745 | -46% | 0 | 0 | — |
case-08 | fail→pass | 14,171 | 7,517 | -47% | 1 | 1 | 0% | 2,349 | 1,586 | -32% | 0 | 0 | — |
case-09 | fail→pass | 8,869 | 2,006 | -77% | 1 | 1 | 0% | 1,444 | 602 | -58% | 0 | 0 | — |
case-10 | fail→pass | 13,288 | 1,695 | -87% | 1 | 1 | 0% | 2,259 | 589 | -74% | 0 | 0 | — |
case-20 | pass→pass | 5,825 | 8,642 | +48% | 1 | 1 | 0% | 1,007 | 1,859 | +85% | 0 | 0 | — |
case-11 | pass→pass | 12,545 | 8,010 | -36% | 1 | 1 | 0% | 2,116 | 1,736 | -18% | 0 | 0 | — |
case-12 | fail→pass | 42,722 | 1,577 | -96% | 1 | 1 | 0% | 7,712 | 565 | -93% | 0 | 0 | — |
case-13 | fail→pass | 10,608 | 1,609 | -85% | 1 | 1 | 0% | 1,828 | 540 | -70% | 0 | 0 | — |
case-14 | fail→pass | 6,854 | 2,191 | -68% | 1 | 1 | 0% | 1,071 | 712 | -34% | 0 | 0 | — |
case-15 | fail→pass | 6,977 | 2,112 | -70% | 1 | 1 | 0% | 1,254 | 739 | -41% | 0 | 0 | — |
case-16 | pass→pass | 9,536 | 3,503 | -63% | 1 | 1 | 0% | 1,623 | 886 | -45% | 0 | 0 | — |
case-17 | fail→pass | 7,552 | 2,674 | -65% | 1 | 1 | 0% | 1,186 | 748 | -37% | 0 | 0 | — |
case-18 | fail→fail | 12,923 | 1,973 | -85% | 1 | 1 | 0% | 2,549 | 707 | -72% | 0 | 0 | — |
case-19 | fail→pass | 6,965 | 1,799 | -74% | 1 | 1 | 0% | 1,148 | 618 | -46% | 0 | 0 | — |
case-23 | fail→pass | 8,362 | 2,073 | -75% | 1 | 1 | 0% | 1,531 | 649 | -58% | 0 | 0 | — |
case-24 | fail→pass | 19,847 | 1,539 | -92% | 1 | 1 | 0% | 3,351 | 536 | -84% | 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. The headline lift of +50 percentage points is the difference between those two pass rates over the 24 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.