Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Voice and audio service management — text-to-speech, voice cloning, and audio generation. Use when an agent needs to generate speech audio, manage voice models, or work with ElevenLabs audio services.
.claude/skills/composio-community-elevenlabs-cli/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-21 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -72% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -19% | 0% |
Voice and audio service management from the terminal.
bashelevenlabs tts --text "Hello world" --voice "Rachel" --output hello.mp3 elevenlabs voices list elevenlabs voices clone --name "MyVoice" --files sample1.mp3 sample2.mp3
ELEVENLABS_API_KEY| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | fail→pass | 18,333 | 11,574 | -37% | 1 | 1 | 0% | 3,340 | 1,703 | -49% | 0 | 0 | — |
case-13 | fail→pass | 12,704 | 2,827 | -78% | 1 | 1 | 0% | 1,856 | 555 | -70% | 0 | 0 | — |
case-14 | fail→pass | 6,741 | 1,845 | -73% | 1 | 1 | 0% | 1,027 | 386 | -62% | 0 | 0 | — |
case-01 | pass→pass | 9,925 | 2,270 | -77% | 1 | 1 | 0% | 1,224 | 408 | -67% | 0 | 0 | — |
case-02 | pass→pass | 6,085 | 2,817 | -54% | 1 | 1 | 0% | 1,087 | 270 | -75% | 0 | 0 | — |
case-03 | fail→pass | 12,469 | 3,089 | -75% | 1 | 1 | 0% | 1,745 | 493 | -72% | 0 | 0 | — |
case-15 | pass→pass | 23,454 | 2,634 | -89% | 1 | 1 | 0% | 3,886 | 438 | -89% | 0 | 0 | — |
case-04 | fail→pass | 89,682 | 30,357 | -66% | 1 | 1 | 0% | 1,957 | 1,591 | -19% | 0 | 0 | — |
case-05 | pass→pass | 8,523 | 2,401 | -72% | 1 | 1 | 0% | 1,484 | 488 | -67% | 0 | 0 | — |
case-06 | fail→pass | 10,461 | 1,653 | -84% | 1 | 1 | 0% | 1,460 | 386 | -74% | 0 | 0 | — |
case-07 | fail→pass | 7,305 | 2,187 | -70% | 1 | 1 | 0% | 1,278 | 341 | -73% | 0 | 0 | — |
case-08 | pass→pass | 3,386 | 2,757 | -19% | 1 | 1 | 0% | 410 | 341 | -17% | 0 | 0 | — |
case-09 | fail→pass | 8,933 | 2,578 | -71% | 1 | 1 | 0% | 1,712 | 440 | -74% | 0 | 0 | — |
case-10 | fail→pass | 5,101 | 37,612 | +637% | 1 | 1 | 0% | 918 | 460 | -50% | 0 | 0 | — |
case-11 | fail→pass | 14,388 | 2,259 | -84% | 1 | 1 | 0% | 2,063 | 505 | -76% | 0 | 0 | — |
case-12 | pass→pass | 13,043 | 4,315 | -67% | 1 | 1 | 0% | 1,809 | 717 | -60% | 0 | 0 | — |
case-16 | pass→pass | 4,857 | 1,918 | -61% | 1 | 1 | 0% | 837 | 400 | -52% | 0 | 0 | — |
case-17 | pass→pass | 5,747 | 1,425 | -75% | 1 | 1 | 0% | 976 | 349 | -64% | 0 | 0 | — |
case-18 | pass→pass | 5,734 | 1,643 | -71% | 1 | 1 | 0% | 700 | 358 | -49% | 0 | 0 | — |
case-19 | pass→pass | 15,656 | 2,618 | -83% | 1 | 1 | 0% | 3,005 | 368 | -88% | 0 | 0 | — |
case-20 | fail→pass | 5,483 | 2,281 | -58% | 1 | 1 | 0% | 1,139 | 565 | -50% | 0 | 0 | — |
case-22 | fail→fail | 13,589 | 8,331 | -39% | 1 | 1 | 0% | 1,850 | 1,497 | -19% | 0 | 0 | — |
case-23 | pass→pass | 36,716 | 9,759 | -73% | 1 | 1 | 0% | 2,800 | 1,768 | -37% | 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 +48 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.