Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Pick the karaoke song that actually fits your voice and the room — so you land it instead of dying on a key change. Use when asked what karaoke song should I sing, pick me a karaoke song, what should I sing for [occasion], or a song for my voice. Produces a few tailored song picks matched to your range and skill, why each works (and the tricky bit to watch), a crowd-pleaser vs a show-off pick, a group/duet option, and a safe fallback for when nerves hit.
.claude/skills/mohitagw15856-karaoke-song-picker/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 13% | 0% |
The wrong karaoke song is a long three minutes — too high, too fast, or too obscure for the room. This picks songs that fit your actual voice and the crowd, warns you about the one hard part, and always keeps a banker in your back pocket for when the confidence wobbles.
Ask for these if not provided:
🎤 Crowd-pleaser: song] — works because familiar/singable]. Watch: the tricky bit]. 🎤 Show-off (if you've got it): song] — shows range/control]. Watch: the hard part]. 🎤 Group/duet: song] — brings people in.
Safe fallback: near-guaranteed win]. Tip: key change / drop the octave / own the chorus].
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 18,118 | 16,738 | -8% | 1 | 1 | 0% | 2,228 | 2,929 | +31% | 0 | 0 | — |
case-01 | fail→pass | 20,728 | 25,223 | +22% | 1 | 1 | 0% | 2,654 | 3,872 | +46% | 0 | 0 | — |
case-02 | fail→pass | 21,054 | 24,377 | +16% | 1 | 1 | 0% | 3,690 | 3,818 | +3% | 0 | 0 | — |
case-03 | fail→pass | 23,077 | 20,415 | -12% | 1 | 1 | 0% | 2,994 | 3,514 | +17% | 0 | 0 | — |
case-04 | pass→pass | 14,792 | 8,340 | -44% | 1 | 1 | 0% | 1,596 | 1,286 | -19% | 0 | 0 | — |
case-05 | fail→pass | 23,142 | 14,723 | -36% | 1 | 1 | 0% | 2,988 | 3,381 | +13% | 0 | 0 | — |
case-06 | pass→pass | 27,504 | 16,362 | -41% | 1 | 1 | 0% | 3,555 | 3,315 | -7% | 0 | 0 | — |
case-08 | fail→fail | 20,278 | 19,464 | -4% | 1 | 1 | 0% | 2,620 | 3,415 | +30% | 0 | 0 | — |
case-09 | fail→fail | 16,631 | 17,007 | +2% | 1 | 1 | 0% | 2,683 | 2,870 | +7% | 0 | 0 | — |
case-10 | pass→pass | 21,202 | 19,668 | -7% | 1 | 1 | 0% | 2,595 | 3,366 | +30% | 0 | 0 | — |
case-11 | fail→pass | 18,562 | 16,633 | -10% | 1 | 1 | 0% | 2,185 | 2,886 | +32% | 0 | 0 | — |
case-12 | pass→pass | 23,598 | 15,156 | -36% | 1 | 1 | 0% | 3,026 | 3,347 | +11% | 0 | 0 | — |
case-13 | pass→pass | 15,367 | 16,871 | +10% | 1 | 1 | 0% | 2,404 | 2,975 | +24% | 0 | 0 | — |
case-14 | fail→pass | 21,776 | 21,170 | -3% | 1 | 1 | 0% | 2,662 | 3,489 | +31% | 0 | 0 | — |
case-15 | fail→pass | 21,265 | 20,392 | -4% | 1 | 1 | 0% | 2,655 | 3,436 | +29% | 0 | 0 | — |
case-16 | fail→pass | 14,146 | 18,636 | +32% | 1 | 1 | 0% | 2,252 | 3,241 | +44% | 0 | 0 | — |
case-17 | pass→pass | 15,631 | 20,935 | +34% | 1 | 1 | 0% | 2,711 | 3,323 | +23% | 0 | 0 | — |
case-18 | pass→pass | 14,416 | 14,757 | +2% | 1 | 1 | 0% | 2,539 | 2,833 | +12% | 0 | 0 | — |
case-19 | fail→pass | 17,786 | 18,605 | +5% | 1 | 1 | 0% | 2,114 | 3,925 | +86% | 0 | 0 | — |
case-20 | pass→pass | 27,037 | 25,145 | -7% | 1 | 1 | 0% | 3,260 | 4,100 | +26% | 0 | 0 | — |
case-21 | pass→fail | 28,124 | 20,816 | -26% | 1 | 1 | 0% | 3,231 | 4,094 | +27% | 0 | 0 | — |
case-22 | pass→pass | 15,905 | 14,420 | -9% | 1 | 1 | 0% | 1,881 | 2,404 | +28% | 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 +41 percentage points is the difference between those two pass rates over the 22 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.