Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →name: x-starship
.claude/skills/x-cmd-starship/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -71% | 0% |
name: x-starship description: | Starship.rs cross-shell prompt with theme management. Minimal, blazing-fast, and infinitely customizable prompt. Auto-downloads starship binary if not available.
Dependency: This is an x-cmd module. Install x-cmd first (see x-cmd skill for installation options). see x-cmd skill for installation.
| Tool | Purpose | Notes | |------|---------|-------| | x-cmd | Required module runtime | brew install x-cmd | | starship | Prompt engine | Optional - x-cmd auto-downloads from GitHub if not present |
Auto-download Security: When starship is not found locally, x-cmd downloads the official binary from GitHub (https://github.com/starship/starship/releases). All downloads are verified with SHA256 checksums before execution.
license: Apache-2.0 compatibility: POSIX Shell
metadata: author: Li Junhao version: "1.0.0" category: x-cmd-extension tags: x-cmd, starship, prompt, theme, shell]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,952 | 5,046 | -61% | 1 | 1 | 0% | 2,008 | 1,128 | -44% | 0 | 0 | — |
case-02 | pass→pass | 11,744 | 5,358 | -54% | 1 | 1 | 0% | 2,074 | 1,038 | -50% | 0 | 0 | — |
case-03 | pass→pass | 14,366 | 32,219 | +124% | 1 | 1 | 0% | 2,251 | 1,616 | -28% | 0 | 0 | — |
case-04 | fail→pass | 15,194 | 17,666 | +16% | 1 | 1 | 0% | 2,399 | 2,459 | +3% | 0 | 0 | — |
case-05 | fail→pass | 12,800 | 2,712 | -79% | 1 | 1 | 0% | 862 | 472 | -45% | 0 | 0 | — |
case-06 | fail→pass | 11,531 | 5,916 | -49% | 1 | 1 | 0% | 1,660 | 1,045 | -37% | 0 | 0 | — |
case-07 | fail→pass | 30,518 | 2,970 | -90% | 1 | 1 | 0% | 2,388 | 697 | -71% | 0 | 0 | — |
case-08 | pass→pass | 11,519 | 13,387 | +16% | 1 | 1 | 0% | 1,891 | 1,238 | -35% | 0 | 0 | — |
case-09 | pass→pass | 13,582 | 15,442 | +14% | 1 | 1 | 0% | 1,903 | 1,667 | -12% | 0 | 0 | — |
case-10 | pass→pass | 20,550 | 3,451 | -83% | 1 | 1 | 0% | 3,572 | 630 | -82% | 0 | 0 | — |
case-11 | fail→pass | 10,091 | 2,621 | -74% | 1 | 1 | 0% | 1,690 | 622 | -63% | 0 | 0 | — |
case-12 | pass→pass | 19,279 | 2,330 | -88% | 1 | 1 | 0% | 1,601 | 536 | -67% | 0 | 0 | — |
case-13 | fail→pass | 15,485 | 5,216 | -66% | 1 | 1 | 0% | 2,680 | 878 | -67% | 0 | 0 | — |
case-14 | fail→pass | 8,362 | 2,425 | -71% | 1 | 1 | 0% | 1,516 | 544 | -64% | 0 | 0 | — |
case-15 | pass→pass | 12,910 | 5,043 | -61% | 1 | 1 | 0% | 1,755 | 1,067 | -39% | 0 | 0 | — |
case-16 | fail→pass | 93,767 | 2,442 | -97% | 1 | 1 | 0% | 1,091 | 532 | -51% | 0 | 0 | — |
case-17 | pass→pass | 29,262 | 9,561 | -67% | 1 | 1 | 0% | 2,524 | 1,781 | -29% | 0 | 0 | — |
case-18 | fail→pass | 20,876 | 2,775 | -87% | 1 | 1 | 0% | 1,360 | 615 | -55% | 0 | 0 | — |
case-19 | fail→pass | 12,060 | 4,445 | -63% | 1 | 1 | 0% | 1,951 | 970 | -50% | 0 | 0 | — |
case-20 | pass→pass | 8,091 | 5,404 | -33% | 1 | 1 | 0% | 1,010 | 1,128 | +12% | 0 | 0 | — |
case-21 | fail→pass | 10,982 | 8,214 | -25% | 1 | 1 | 0% | 1,927 | 1,592 | -17% | 0 | 0 | — |
case-22 | pass→pass | 8,100 | 12,198 | +51% | 1 | 1 | 0% | 1,189 | 1,153 | -3% | 0 | 0 | — |
case-23 | pass→pass | 10,668 | 5,903 | -45% | 1 | 1 | 0% | 1,482 | 1,060 | -28% | 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 +52 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.