Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Takes a scenario or decision and simulates multiple outcome branches. Uses Monte Carlo tree search to explore the decision space, evaluating each branch on probability and impact. Returns a ranked set of outcomes with recommended actions.
.claude/skills/miosa-osa-simulate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -80% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -84% | 0% |
Monte Carlo scenario planning and impact analysis.
Runs Monte Carlo Tree Search for scenario planning. Evaluates potential decisions by simulating outcomes across multiple paths. Reports expected value, risk distribution, and recommended actions.
/simulate "What if we price AI Masters at $2K per seat?"
/simulate "Impact of hiring two more devs" --scenarios 100| Flag | Description | Default | |------|-------------|---------| | --scenarios | Number of simulation paths | 50 |
bashcd engine && mix optimal.simulate "scenario description" cd engine && mix optimal.impact "decision description"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | pass→pass | 8,668 | 15,309 | +77% | 1 | 1 | 0% | 1,564 | 2,251 | +44% | 0 | 0 | — |
case-01 | fail→pass | 22,123 | 15,174 | -31% | 1 | 1 | 0% | 3,627 | 2,387 | -34% | 0 | 0 | — |
case-02 | fail→fail | 26,969 | 17,988 | -33% | 1 | 1 | 0% | 4,184 | 3,003 | -28% | 0 | 0 | — |
case-03 | fail→pass | 24,867 | 15,773 | -37% | 1 | 1 | 0% | 4,301 | 2,818 | -34% | 0 | 0 | — |
case-04 | fail→pass | 6,526 | 3,432 | -47% | 1 | 1 | 0% | 1,090 | 603 | -45% | 0 | 0 | — |
case-05 | fail→pass | 11,553 | 1,366 | -88% | 1 | 1 | 0% | 1,767 | 352 | -80% | 0 | 0 | — |
case-06 | fail→pass | 16,342 | 2,021 | -88% | 1 | 1 | 0% | 2,896 | 471 | -84% | 0 | 0 | — |
case-07 | fail→pass | 14,380 | 2,991 | -79% | 1 | 1 | 0% | 1,074 | 510 | -53% | 0 | 0 | — |
case-08 | pass→pass | 14,971 | 11,391 | -24% | 1 | 1 | 0% | 2,291 | 1,935 | -16% | 0 | 0 | — |
case-09 | fail→pass | 9,939 | 2,150 | -78% | 1 | 1 | 0% | 1,311 | 429 | -67% | 0 | 0 | — |
case-10 | fail→pass | 4,478 | 2,050 | -54% | 1 | 1 | 0% | 754 | 400 | -47% | 0 | 0 | — |
case-11 | fail→pass | 9,100 | 4,775 | -48% | 1 | 1 | 0% | 1,338 | 813 | -39% | 0 | 0 | — |
case-12 | pass→pass | 11,864 | 3,424 | -71% | 1 | 1 | 0% | 1,803 | 666 | -63% | 0 | 0 | — |
case-13 | fail→pass | 11,938 | 2,472 | -79% | 1 | 1 | 0% | 1,973 | 469 | -76% | 0 | 0 | — |
case-14 | fail→pass | 11,641 | 2,411 | -79% | 1 | 1 | 0% | 1,558 | 497 | -68% | 0 | 0 | — |
case-15 | fail→pass | 5,927 | 1,938 | -67% | 1 | 1 | 0% | 709 | 387 | -45% | 0 | 0 | — |
case-16 | pass→pass | 12,548 | 2,593 | -79% | 1 | 1 | 0% | 1,817 | 580 | -68% | 0 | 0 | — |
case-17 | fail→pass | 7,494 | 3,052 | -59% | 1 | 1 | 0% | 1,143 | 674 | -41% | 0 | 0 | — |
case-18 | fail→pass | 22,732 | 5,302 | -77% | 1 | 1 | 0% | 3,558 | 856 | -76% | 0 | 0 | — |
case-20 | pass→pass | 8,324 | 4,470 | -46% | 1 | 1 | 0% | 1,445 | 930 | -36% | 0 | 0 | — |
case-21 | pass→pass | 8,740 | 8,257 | -6% | 1 | 1 | 0% | 1,679 | 1,850 | +10% | 0 | 0 | — |
case-22 | pass→pass | 8,967 | 6,288 | -30% | 1 | 1 | 0% | 1,264 | 1,031 | -18% | 0 | 0 | — |
case-23 | fail→pass | 15,088 | 2,075 | -86% | 1 | 1 | 0% | 2,582 | 400 | -85% | 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. 23 cases were attempted. The headline lift of +65 percentage points is the difference between those two pass rates over the 23 comparable cases.
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.