Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Render videos, audio, and still images programmatically with React using terminal commands. Use when an agent needs to preview compositions, list available compositions, or render media artifacts from a Remotion project.
.claude/skills/composio-community-remotion-cli/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -32% | 0% |
Official CLI for Remotion, a React-based framework for making videos programmatically.
bash# Run without installing globally npx remotion --help # Or install in a project npm install remotion
Most usage is via npx remotion ... inside a Remotion project.
bash# Start the Remotion Studio preview UI npx remotion studio # List available composition IDs npx remotion compositions # Render a composition to a file npx remotion render <composition-id> out/video.mp4 # Create a bundle for rendering workflows npx remotion bundle
npx remotionstudio is useful for previewing and iterating locallyrender is the main command for automated media generationcompositions helps agents discover which composition IDs exist before rendering| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→pass | 5,608 | 1,609 | -71% | 1 | 1 | 0% | 1,056 | 556 | -47% | 0 | 0 | — |
case-01 | fail→pass | 5,062 | 4,626 | -9% | 1 | 1 | 0% | 774 | 688 | -11% | 0 | 0 | — |
case-02 | pass→pass | 9,840 | 3,970 | -60% | 1 | 1 | 0% | 1,697 | 964 | -43% | 0 | 0 | — |
case-04 | pass→pass | 7,771 | 5,736 | -26% | 1 | 1 | 0% | 1,165 | 998 | -14% | 0 | 0 | — |
case-05 | pass→pass | 11,488 | 4,805 | -58% | 1 | 1 | 0% | 1,952 | 1,159 | -41% | 0 | 0 | — |
case-06 | pass→pass | 3,765 | 2,251 | -40% | 1 | 1 | 0% | 570 | 643 | +13% | 0 | 0 | — |
case-07 | fail→pass | 4,567 | 3,058 | -33% | 1 | 1 | 0% | 665 | 618 | -7% | 0 | 0 | — |
case-08 | pass→pass | 10,249 | 3,564 | -65% | 1 | 1 | 0% | 1,658 | 858 | -48% | 0 | 0 | — |
case-09 | pass→pass | 4,562 | 1,582 | -65% | 1 | 1 | 0% | 789 | 568 | -28% | 0 | 0 | — |
case-10 | pass→pass | 12,569 | 9,129 | -27% | 1 | 1 | 0% | 2,108 | 1,463 | -31% | 0 | 0 | — |
case-11 | pass→pass | 6,005 | 2,732 | -55% | 1 | 1 | 0% | 713 | 661 | -7% | 0 | 0 | — |
case-12 | fail→pass | 7,043 | 5,075 | -28% | 1 | 1 | 0% | 1,339 | 884 | -34% | 0 | 0 | — |
case-13 | pass→pass | 11,102 | 17,459 | +57% | 1 | 1 | 0% | 1,692 | 2,787 | +65% | 0 | 0 | — |
case-14 | fail→pass | 7,230 | 2,888 | -60% | 1 | 1 | 0% | 1,206 | 821 | -32% | 0 | 0 | — |
case-15 | fail→fail | 7,525 | 5,323 | -29% | 1 | 1 | 0% | 717 | 1,006 | +40% | 0 | 0 | — |
case-16 | pass→pass | 10,184 | 7,726 | -24% | 1 | 1 | 0% | 1,632 | 1,138 | -30% | 0 | 0 | — |
case-17 | pass→pass | 3,747 | 3,236 | -14% | 1 | 1 | 0% | 628 | 785 | +25% | 0 | 0 | — |
case-18 | pass→pass | 15,400 | 3,421 | -78% | 1 | 1 | 0% | 1,225 | 731 | -40% | 0 | 0 | — |
case-19 | pass→pass | 3,501 | 3,860 | +10% | 1 | 1 | 0% | 503 | 695 | +38% | 0 | 0 | — |
case-20 | pass→pass | 2,903 | 3,016 | +4% | 1 | 1 | 0% | 597 | 958 | +60% | 0 | 0 | — |
case-21 | pass→pass | 7,437 | 5,175 | -30% | 1 | 1 | 0% | 1,141 | 1,268 | +11% | 0 | 0 | — |
case-22 | pass→pass | 5,191 | 5,051 | -3% | 1 | 1 | 0% | 1,016 | 1,314 | +29% | 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. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 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.