Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generates terminal recordings using VHS tape scripts and produces GIF outputs. Use when creating demo GIFs or documenting CLI workflows for tutorials.
.claude/skills/athola-vhs-recording/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -51% | 0% |
Generate professional terminal recordings from VHS tape files.
VHS converts declarative tape files into animated GIFs of terminal sessions. Tape files define commands, timing, and terminal appearance.
- Locate and validate tape file
- Check VHS installation status
- Execute VHS recording
- Verify output GIF creationmodules/tape-syntax.md for VHS tape file directivesmodules/execution.md for recording workflow detailsOutput directive specifies GIF destinationbashwhich vhs && vhs --version
If not installed:
bash# Linux/WSL go install github.com/charmbracelet/vhs@latest # macOS brew install charmbracelet/tap/vhs # Also requires ttyd and ffmpeg
bashvhs <tape-file.tape>
VHS will:
If vhs is not found, verify that your Go bin directory is in your PATH (typically ~/go/bin). If the recording fails to start, ensure ttyd and ffmpeg are installed, as VHS depends on them for terminal emulation and video encoding. For "permission denied" errors when writing the GIF, check that the output directory exists and is writable.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,385 | 3,447 | -36% | 1 | 1 | 0% | 195 | 813 | +317% | 0 | 0 | — |
case-02 | fail→fail | 5,889 | 3,705 | -37% | 1 | 1 | 0% | 220 | 951 | +332% | 0 | 0 | — |
case-03 | fail→fail | 5,278 | 6,123 | +16% | 1 | 1 | 0% | 175 | 812 | +364% | 0 | 0 | — |
case-04 | fail→pass | 3,410 | 1,987 | -42% | 1 | 1 | 0% | 500 | 825 | +65% | 0 | 0 | — |
case-05 | pass→pass | 4,344 | 2,067 | -52% | 1 | 1 | 0% | 643 | 810 | +26% | 0 | 0 | — |
case-06 | fail→pass | 4,187 | 1,898 | -55% | 1 | 1 | 0% | 558 | 784 | +41% | 0 | 0 | — |
case-07 | fail→pass | 11,064 | 3,094 | -72% | 1 | 1 | 0% | 1,719 | 931 | -46% | 0 | 0 | — |
case-08 | pass→pass | 3,097 | 1,660 | -46% | 1 | 1 | 0% | 477 | 724 | +52% | 0 | 0 | — |
case-09 | pass→pass | 8,853 | 2,065 | -77% | 1 | 1 | 0% | 1,321 | 802 | -39% | 0 | 0 | — |
case-10 | pass→pass | 2,857 | 2,254 | -21% | 1 | 1 | 0% | 412 | 823 | +100% | 0 | 0 | — |
case-11 | pass→pass | 13,681 | 6,024 | -56% | 1 | 1 | 0% | 2,293 | 1,492 | -35% | 0 | 0 | — |
case-12 | pass→pass | 6,122 | 3,787 | -38% | 1 | 1 | 0% | 1,106 | 1,213 | +10% | 0 | 0 | — |
case-13 | pass→pass | 10,822 | 3,545 | -67% | 1 | 1 | 0% | 1,882 | 1,033 | -45% | 0 | 0 | — |
case-18 | fail→pass | 8,479 | 1,925 | -77% | 1 | 1 | 0% | 1,318 | 769 | -42% | 0 | 0 | — |
case-14 | pass→pass | 6,358 | 1,807 | -72% | 1 | 1 | 0% | 915 | 870 | -5% | 0 | 0 | — |
case-15 | pass→fail | 9,546 | 3,314 | -65% | 1 | 1 | 0% | 1,532 | 1,100 | -28% | 0 | 0 | — |
case-16 | fail→pass | 13,883 | 3,128 | -77% | 1 | 1 | 0% | 2,097 | 1,024 | -51% | 0 | 0 | — |
case-17 | fail→pass | 10,503 | 1,605 | -85% | 1 | 1 | 0% | 1,466 | 736 | -50% | 0 | 0 | — |
case-19 | fail→pass | 9,852 | 2,340 | -76% | 1 | 1 | 0% | 1,665 | 740 | -56% | 0 | 0 | — |
case-20 | pass→pass | 15,891 | 2,855 | -82% | 1 | 1 | 0% | 2,576 | 901 | -65% | 0 | 0 | — |
case-21 | pass→pass | 9,726 | 3,121 | -68% | 1 | 1 | 0% | 1,440 | 967 | -33% | 0 | 0 | — |
case-22 | pass→pass | 9,954 | 4,674 | -53% | 1 | 1 | 0% | 1,525 | 1,132 | -26% | 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 19 counted toward the lift figure. The other 3 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 +27 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.