Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Takes content in one genre and re-encodes it in another while preserving the core signal. A spec becomes a brief. A transcript becomes action items. A brain dump becomes a plan. Matches the target genre's skeleton and the receiver's bandwidth.
.claude/skills/miosa-osa-translate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 826% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 110% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 576% | 0% |
> Convert content between genres while preserving the core signal.
/translate <file> --from <genre> --to <genre> [--for <person>]Takes content in one genre and re-encodes it in another while preserving the core signal. A spec becomes a brief. A transcript becomes action items. A brain dump becomes a plan. Matches the target genre's skeleton and the receiver's bandwidth.
--for specified, match that person's preferred genre and bandwidth.bash# Convert a spec into a brief for a salesperson /translate docs/system-spec.md --from spec --to brief --for "sales team" # Convert a transcript into action items /translate signals/2026-03-18-jordan-debrief.md --from transcript --to plan # Convert a brain dump into a structured plan /translate rhythm/weekly-dump.md --from note --to plan
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,097 | 4,148 | -32% | 1 | 1 | 0% | 973 | 939 | -3% | 0 | 0 | — |
case-02 | pass→pass | 11,571 | 9,070 | -22% | 1 | 1 | 0% | 1,781 | 1,715 | -4% | 0 | 0 | — |
case-03 | fail→pass | 6,594 | 14,226 | +116% | 1 | 1 | 0% | 290 | 2,686 | +826% | 0 | 0 | — |
case-04 | fail→fail | 5,878 | 5,242 | -11% | 1 | 1 | 0% | 978 | 1,207 | +23% | 0 | 0 | — |
case-05 | fail→pass | 11,541 | 2,286 | -80% | 1 | 1 | 0% | 1,898 | 708 | -63% | 0 | 0 | — |
case-06 | fail→pass | 11,661 | 19,744 | +69% | 1 | 1 | 0% | 1,899 | 2,889 | +52% | 0 | 0 | — |
case-07 | pass→pass | 22,109 | 15,247 | -31% | 1 | 1 | 0% | 3,598 | 2,827 | -21% | 0 | 0 | — |
case-08 | fail→fail | 10,381 | 7,302 | -30% | 1 | 1 | 0% | 1,377 | 1,357 | -1% | 0 | 0 | — |
case-09 | fail→pass | 12,048 | 16,572 | +38% | 1 | 1 | 0% | 1,423 | 2,990 | +110% | 0 | 0 | — |
case-10 | fail→pass | 4,753 | 10,361 | +118% | 1 | 1 | 0% | 285 | 1,926 | +576% | 0 | 0 | — |
case-11 | pass→pass | 20,543 | 11,443 | -44% | 1 | 1 | 0% | 3,187 | 2,053 | -36% | 0 | 0 | — |
case-12 | fail→pass | 24,207 | 15,277 | -37% | 1 | 1 | 0% | 3,778 | 2,785 | -26% | 0 | 0 | — |
case-13 | fail→fail | 11,652 | 15,179 | +30% | 1 | 1 | 0% | 1,790 | 2,743 | +53% | 0 | 0 | — |
case-14 | fail→fail | 4,972 | 7,231 | +45% | 1 | 1 | 0% | 331 | 1,510 | +356% | 0 | 0 | — |
case-15 | fail→pass | 4,098 | 13,431 | +228% | 1 | 1 | 0% | 678 | 2,062 | +204% | 0 | 0 | — |
case-16 | pass→pass | 9,476 | 11,730 | +24% | 1 | 1 | 0% | 1,474 | 2,083 | +41% | 0 | 0 | — |
case-17 | fail→fail | 33,113 | 10,505 | -68% | 1 | 1 | 0% | 4,831 | 2,074 | -57% | 0 | 0 | — |
case-18 | fail→fail | 12,456 | 4,450 | -64% | 1 | 1 | 0% | 1,846 | 521 | -72% | 0 | 0 | — |
case-19 | fail→pass | 3,681 | 10,571 | +187% | 1 | 1 | 0% | 588 | 2,002 | +240% | 0 | 0 | — |
case-20 | fail→fail | 1,775 | 6,099 | +244% | 1 | 1 | 0% | 205 | 1,345 | +556% | 0 | 0 | — |
case-21 | fail→fail | 1,219 | 3,791 | +211% | 1 | 1 | 0% | 186 | 890 | +378% | 0 | 0 | — |
case-22 | fail→fail | 1,421 | 4,848 | +241% | 1 | 1 | 0% | 242 | 996 | +312% | 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 18 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 +36 percentage points is the difference between those two pass rates over the 18 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.