Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Maps the ripple effects of a proposed change across the knowledge base. Identifies which nodes, people, revenue streams, and processes are affected. Classifies impacts as direct (1st order) or indirect (2nd/3rd order).
.claude/skills/miosa-osa-impact/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -17% | 0% |
> Analyze the downstream impact of a change across all nodes.
/impact "<change>" [--scope <node>]Maps the ripple effects of a proposed change across the knowledge base. Identifies which nodes, people, revenue streams, and processes are affected. Classifies impacts as direct (1st order) or indirect (2nd/3rd order).
Runs: cd engine && mix optimal.impact
Process:
bash# Impact of a team change /impact "Lead developer takes 2 weeks off" # Impact of a pricing change /impact "Raise enterprise pricing to $5K/mo" # Scoped impact analysis /impact "Switch from Firecracker to containers" --scope platform
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 28,651 | 20,044 | -30% | 1 | 1 | 0% | 3,814 | 2,935 | -23% | 0 | 0 | — |
case-02 | fail→fail | 29,900 | 17,426 | -42% | 1 | 1 | 0% | 3,911 | 2,907 | -26% | 0 | 0 | — |
case-08 | pass→pass | 23,226 | 16,020 | -31% | 1 | 1 | 0% | 2,854 | 2,772 | -3% | 0 | 0 | — |
case-03 | fail→pass | 18,769 | 16,704 | -11% | 1 | 1 | 0% | 2,650 | 2,782 | +5% | 0 | 0 | — |
case-04 | fail→pass | 17,969 | 18,006 | +0% | 1 | 1 | 0% | 2,833 | 3,154 | +11% | 0 | 0 | — |
case-05 | fail→pass | 15,520 | 4,413 | -72% | 1 | 1 | 0% | 2,338 | 958 | -59% | 0 | 0 | — |
case-06 | fail→fail | 18,335 | 2,015 | -89% | 1 | 1 | 0% | 1,099 | 595 | -46% | 0 | 0 | — |
case-07 | fail→pass | 18,167 | 14,228 | -22% | 1 | 1 | 0% | 2,885 | 2,405 | -17% | 0 | 0 | — |
case-09 | fail→pass | 15,865 | 13,185 | -17% | 1 | 1 | 0% | 2,319 | 1,945 | -16% | 0 | 0 | — |
case-10 | fail→pass | 30,328 | 31,643 | +4% | 1 | 1 | 0% | 3,799 | 4,581 | +21% | 0 | 0 | — |
case-11 | fail→pass | 18,183 | 19,020 | +5% | 1 | 1 | 0% | 2,662 | 3,064 | +15% | 0 | 0 | — |
case-12 | fail→fail | 19,994 | 21,848 | +9% | 1 | 1 | 0% | 2,987 | 2,958 | -1% | 0 | 0 | — |
case-13 | pass→pass | 17,546 | 15,000 | -15% | 1 | 1 | 0% | 2,666 | 2,731 | +2% | 0 | 0 | — |
case-14 | fail→pass | 19,923 | 69,082 | +247% | 1 | 1 | 0% | 3,060 | 1,881 | -39% | 0 | 0 | — |
case-15 | pass→pass | 12,824 | 5,651 | -56% | 1 | 1 | 0% | 2,029 | 1,144 | -44% | 0 | 0 | — |
case-16 | fail→pass | 19,221 | 18,060 | -6% | 1 | 1 | 0% | 3,350 | 3,015 | -10% | 0 | 0 | — |
case-17 | fail→pass | 20,755 | 25,879 | +25% | 1 | 1 | 0% | 3,186 | 3,158 | -1% | 0 | 0 | — |
case-18 | fail→pass | 16,940 | 15,619 | -8% | 1 | 1 | 0% | 2,359 | 2,577 | +9% | 0 | 0 | — |
case-19 | fail→fail | 19,273 | 24,061 | +25% | 1 | 1 | 0% | 3,014 | 3,438 | +14% | 0 | 0 | — |
case-20 | pass→pass | 6,323 | 9,516 | +50% | 1 | 1 | 0% | 1,329 | 1,553 | +17% | 0 | 0 | — |
case-21 | pass→pass | 19,021 | 15,360 | -19% | 1 | 1 | 0% | 2,876 | 3,109 | +8% | 0 | 0 | — |
case-22 | pass→pass | 15,151 | 19,609 | +29% | 1 | 1 | 0% | 3,055 | 3,376 | +11% | 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. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 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 +55 percentage points is the difference between those two pass rates over the 21 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.