Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Install, update, trust, or inspect DeepSeek skills from GitHub or local skill folders. Use when the user asks for available skills or wants a community skill installed.
.claude/skills/hmbown-skill-installer/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 7 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -29% | 0% |
Use this skill when the user wants to find, install, update, trust, or remove a Codewhale skill.
/skills/skill <name>/skill new/skill install github:<owner>/<repo>/skill update/skill uninstall <name>/skill trust <name>SKILL.md/skill commands over ad hoc copying.SKILL.md before recommendingtrust. Treat scripts and companion files as untrusted until reviewed.
appear immediately in the available-skills list.
skill directories take precedence over global ~/.codewhale/skills, with
Do not execute community skill scripts unless the user explicitly asks and the skill has been reviewed or trusted.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | pass→pass | 10,344 | 11,136 | +8% | 1 | 1 | 0% | 1,458 | 677 | -54% | 0 | 0 | — |
case-14 | fail→pass | 15,844 | 13,220 | -17% | 1 | 1 | 0% | 1,578 | 661 | -58% | 0 | 0 | — |
case-01 | fail→pass | 19,569 | 12,506 | -36% | 1 | 1 | 0% | 2,539 | 1,174 | -54% | 0 | 0 | — |
case-02 | fail→pass | 14,352 | 12,400 | -14% | 1 | 1 | 0% | 2,251 | 1,494 | -34% | 0 | 0 | — |
case-03 | fail→pass | 19,786 | 14,502 | -27% | 1 | 1 | 0% | 2,346 | 1,205 | -49% | 0 | 0 | — |
case-04 | fail→pass | 8,708 | 3,330 | -62% | 1 | 1 | 0% | 1,068 | 755 | -29% | 0 | 0 | — |
case-05 | fail→pass | 22,786 | 7,508 | -67% | 1 | 1 | 0% | 1,828 | 774 | -58% | 0 | 0 | — |
case-06 | fail→pass | 16,452 | 5,016 | -70% | 1 | 1 | 0% | 1,641 | 944 | -42% | 0 | 0 | — |
case-07 | fail→pass | 16,441 | 39,175 | +138% | 1 | 1 | 0% | 1,865 | 1,107 | -41% | 0 | 0 | — |
case-08 | pass→pass | 18,882 | 9,889 | -48% | 1 | 1 | 0% | 2,094 | 943 | -55% | 0 | 0 | — |
case-10 | pass→pass | 11,626 | 3,156 | -73% | 1 | 1 | 0% | 748 | 602 | -20% | 0 | 0 | — |
case-11 | fail→pass | 10,278 | 7,931 | -23% | 1 | 1 | 0% | 570 | 650 | +14% | 0 | 0 | — |
case-12 | fail→pass | 19,185 | 8,355 | -56% | 1 | 1 | 0% | 2,013 | 732 | -64% | 0 | 0 | — |
case-13 | fail→pass | 12,895 | 9,185 | -29% | 1 | 1 | 0% | 1,677 | 786 | -53% | 0 | 0 | — |
case-15 | fail→pass | 7,500 | 8,691 | +16% | 1 | 1 | 0% | 904 | 589 | -35% | 0 | 0 | — |
case-16 | fail→pass | 12,272 | 13,172 | +7% | 1 | 1 | 0% | 1,068 | 590 | -45% | 0 | 0 | — |
case-17 | fail→pass | 14,558 | 7,919 | -46% | 1 | 1 | 0% | 1,340 | 602 | -55% | 0 | 0 | — |
case-18 | pass→pass | 14,746 | 8,500 | -42% | 1 | 1 | 0% | 1,343 | 541 | -60% | 0 | 0 | — |
case-19 | pass→pass | 18,208 | 18,755 | +3% | 1 | 1 | 0% | 1,795 | 636 | -65% | 0 | 0 | — |
case-20 | fail→pass | 11,866 | 2,599 | -78% | 1 | 1 | 0% | 901 | 569 | -37% | 0 | 0 | — |
case-21 | pass→pass | 12,409 | 5,883 | -53% | 1 | 1 | 0% | 1,936 | 1,183 | -39% | 0 | 0 | — |
case-22 | pass→pass | 39,114 | 19,738 | -50% | 1 | 1 | 0% | 3,406 | 2,944 | -14% | 0 | 0 | — |
case-23 | pass→pass | 16,675 | 12,131 | -27% | 1 | 1 | 0% | 1,583 | 1,379 | -13% | 0 | 0 | — |
case-24 | pass→pass | 9,661 | 9,401 | -3% | 1 | 1 | 0% | 667 | 786 | +18% | 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 +63 percentage points is the difference between those two pass rates over the 24 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/11/2026 | +57% |
| gemini-3.6-flash | verified | 8/7/2026 | +73% |
Other measured skills in the registry, with their headline benchmark lift.