Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create or edit the HyperFrames project owned by an active iPolloWork Video Studio session without changing the embedded Studio, starting another server, or writing to another project.
.claude/skills/devin-axis-ipollowork-video-studio/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 255% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -34% | 0% |
Use this Skill only for the video project owned by the active iPolloWork session. The built-in Video Studio, timeline, preview, templates, and HyperFrames runtime exist independently of this Skill.
index.html path as authoritative.index.html, confirmed brief, template metadata, and design-tokens.css before editing.video/<session-id>/ project.--ipw-* design tokens when the project provides them.index.html and keep referenced assets inside the same project.If the active session provides stricter timing, template, media, or validation instructions, those instructions take precedence.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 50,304 | 14,167 | -72% | 1 | 1 | 0% | 8,230 | 516 | -94% | 0 | 0 | — |
case-06 | fail→fail | 18,704 | 16,552 | -12% | 1 | 1 | 0% | 2,759 | 2,766 | +0% | 0 | 0 | — |
case-02 | fail→fail | 10,357 | 15,052 | +45% | 1 | 1 | 0% | 731 | 627 | -14% | 0 | 0 | — |
case-03 | fail→fail | 25,577 | 14,077 | -45% | 1 | 1 | 0% | 2,525 | 538 | -79% | 0 | 0 | — |
case-04 | fail→pass | 13,276 | 8,344 | -37% | 1 | 1 | 0% | 257 | 912 | +255% | 0 | 0 | — |
case-05 | fail→pass | 10,732 | 9,104 | -15% | 1 | 1 | 0% | 1,004 | 993 | -1% | 0 | 0 | — |
case-07 | pass→pass | 11,132 | 8,069 | -28% | 1 | 1 | 0% | 968 | 876 | -10% | 0 | 0 | — |
case-08 | fail→pass | 14,597 | 8,328 | -43% | 1 | 1 | 0% | 1,415 | 854 | -40% | 0 | 0 | — |
case-09 | fail→fail | 19,880 | 16,107 | -19% | 1 | 1 | 0% | 2,413 | 2,259 | -6% | 0 | 0 | — |
case-10 | fail→pass | 14,428 | 9,268 | -36% | 1 | 1 | 0% | 1,388 | 1,038 | -25% | 0 | 0 | — |
case-11 | pass→pass | 16,703 | 9,969 | -40% | 1 | 1 | 0% | 1,734 | 1,200 | -31% | 0 | 0 | — |
case-12 | pass→pass | 17,630 | 14,312 | -19% | 1 | 1 | 0% | 2,002 | 1,804 | -10% | 0 | 0 | — |
case-13 | fail→pass | 16,255 | 10,183 | -37% | 1 | 1 | 0% | 1,741 | 1,144 | -34% | 0 | 0 | — |
case-14 | fail→fail | 30,027 | 8,262 | -72% | 1 | 1 | 0% | 1,514 | 785 | -48% | 0 | 0 | — |
case-15 | pass→pass | 10,696 | 7,493 | -30% | 1 | 1 | 0% | 919 | 763 | -17% | 0 | 0 | — |
case-16 | fail→fail | 10,008 | 7,858 | -21% | 1 | 1 | 0% | 776 | 895 | +15% | 0 | 0 | — |
case-17 | pass→pass | 15,051 | 11,189 | -26% | 1 | 1 | 0% | 1,702 | 1,316 | -23% | 0 | 0 | — |
case-18 | fail→pass | 19,149 | 14,911 | -22% | 1 | 1 | 0% | 2,165 | 2,051 | -5% | 0 | 0 | — |
case-19 | fail→pass | 14,244 | 9,467 | -34% | 1 | 1 | 0% | 1,336 | 1,079 | -19% | 0 | 0 | — |
case-20 | pass→pass | 17,858 | 11,385 | -36% | 1 | 1 | 0% | 1,959 | 1,530 | -22% | 0 | 0 | — |
case-21 | pass→pass | 13,083 | 7,476 | -43% | 1 | 1 | 0% | 1,345 | 765 | -43% | 0 | 0 | — |
case-22 | pass→pass | 11,457 | 9,731 | -15% | 1 | 1 | 0% | 1,017 | 1,098 | +8% | 0 | 0 | — |
case-23 | pass→pass | 16,029 | 12,735 | -21% | 1 | 1 | 0% | 1,801 | 1,528 | -15% | 0 | 0 | — |
case-24 | pass→pass | 20,259 | 9,388 | -54% | 1 | 1 | 0% | 2,186 | 1,087 | -50% | 0 | 0 | — |
case-25 | fail→pass | 14,849 | 9,380 | -37% | 1 | 1 | 0% | 1,545 | 1,060 | -31% | 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. 25 cases were attempted, and 21 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 +32 percentage points is the difference between those two pass rates over the 21 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.