Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fetch transcripts, metadata, and captions from YouTube videos. Use when you need to summarize a video, answer questions about its content, extract key points, compare videos, or pull quotes with timestamps. Supports auto-generated and manual captions, multiple languages, and translation.
.claude/skills/cowork-os-youtube/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✓→✗ | ▼ Worse | 5% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | -33% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -33% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 8% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -7% | 0% |
Fetch transcripts, metadata, and captions from YouTube videos. Use when you need to summarize a video, answer questions about its content, extract key points, compare videos, or pull quotes with timestamps. Supports auto-generated and manual captions, multiple languages, and translation.
| Name | Type | Required | Description | |---|---|---|---| | url | string | No | YouTube video URL or video ID |
../youtube.json.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | pass→pass | 12,409 | 6,191 | -50% | 1 | 1 | 0% | 2,312 | 1,540 | -33% | 0 | 0 | — |
case-01 | fail→fail | 5,824 | 5,246 | -10% | 1 | 1 | 0% | 858 | 542 | -37% | 0 | 0 | — |
case-02 | fail→fail | 6,254 | 17,009 | +172% | 1 | 1 | 0% | 915 | 1,433 | +57% | 0 | 0 | — |
case-03 | pass→pass | 13,978 | 7,694 | -45% | 1 | 1 | 0% | 2,726 | 1,836 | -33% | 0 | 0 | — |
case-04 | pass→pass | 5,324 | 3,938 | -26% | 1 | 1 | 0% | 877 | 948 | +8% | 0 | 0 | — |
case-05 | pass→pass | 7,047 | 4,319 | -39% | 1 | 1 | 0% | 1,187 | 1,105 | -7% | 0 | 0 | — |
case-06 | pass→pass | 8,642 | 5,518 | -36% | 1 | 1 | 0% | 1,450 | 1,279 | -12% | 0 | 0 | — |
case-08 | pass→pass | 5,322 | 3,466 | -35% | 1 | 1 | 0% | 863 | 914 | +6% | 0 | 0 | — |
case-09 | pass→pass | 10,772 | 6,669 | -38% | 1 | 1 | 0% | 1,830 | 1,519 | -17% | 0 | 0 | — |
case-10 | pass→pass | 12,075 | 7,505 | -38% | 1 | 1 | 0% | 2,235 | 1,752 | -22% | 0 | 0 | — |
case-11 | pass→pass | 10,706 | 7,574 | -29% | 1 | 1 | 0% | 2,072 | 1,825 | -12% | 0 | 0 | — |
case-12 | pass→pass | 7,162 | 6,901 | -4% | 1 | 1 | 0% | 1,175 | 1,422 | +21% | 0 | 0 | — |
case-13 | fail→fail | 13,789 | 6,904 | -50% | 1 | 1 | 0% | 2,618 | 1,642 | -37% | 0 | 0 | — |
case-14 | pass→pass | 10,038 | 5,496 | -45% | 1 | 1 | 0% | 1,643 | 1,280 | -22% | 0 | 0 | — |
case-15 | fail→fail | 6,351 | 5,583 | -12% | 1 | 1 | 0% | 1,106 | 1,352 | +22% | 0 | 0 | — |
case-16 | pass→fail | 11,467 | 10,027 | -13% | 1 | 1 | 0% | 2,102 | 2,212 | +5% | 0 | 0 | — |
case-17 | pass→pass | 3,477 | 3,078 | -11% | 1 | 1 | 0% | 524 | 836 | +60% | 0 | 0 | — |
case-18 | pass→pass | 9,424 | 6,834 | -27% | 1 | 1 | 0% | 1,503 | 1,504 | +0% | 0 | 0 | — |
case-19 | pass→pass | 3,748 | 3,882 | +4% | 1 | 1 | 0% | 680 | 991 | +46% | 0 | 0 | — |
case-20 | pass→pass | 10,307 | 6,507 | -37% | 1 | 1 | 0% | 1,793 | 1,424 | -21% | 0 | 0 | — |
case-21 | pass→pass | 11,369 | 7,591 | -33% | 1 | 1 | 0% | 1,900 | 1,601 | -16% | 0 | 0 | — |
case-22 | pass→pass | 9,807 | 8,082 | -18% | 1 | 1 | 0% | 1,666 | 1,765 | +6% | 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, and 21 counted toward the lift figure. The other 1 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 -100 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.