Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate alternatives for every known approach — ensure no approach goes unchallenged.
.claude/skills/yogsoth-ai-alternatives-generation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✓→✓ | ＝ Same ✓ | 42% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 35% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 61% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | 27% | 0% |
Generate alternatives for every known approach.
Subagent — spawned via subagent-spawning/spawn-agent skill.
Alternatives generation requires systematically considering each known approach and producing genuinely different alternatives (not variations). Benefits from dedicated context that can maintain the "different, not better" mindset.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 20,801 | 34,716 | +67% | 1 | 1 | 0% | 2,954 | 4,761 | +61% | 0 | 0 | — |
case-02 | fail→fail | 20,050 | 25,736 | +28% | 1 | 1 | 0% | 3,062 | 3,893 | +27% | 0 | 0 | — |
case-03 | fail→fail | 20,448 | 23,730 | +16% | 1 | 1 | 0% | 3,001 | 2,801 | -7% | 0 | 0 | — |
case-04 | pass→pass | 20,059 | 27,666 | +38% | 1 | 1 | 0% | 2,897 | 4,127 | +42% | 0 | 0 | — |
case-05 | pass→pass | 20,199 | 27,760 | +37% | 1 | 1 | 0% | 3,144 | 4,253 | +35% | 0 | 0 | — |
case-06 | pass→pass | 32,977 | 29,651 | -10% | 1 | 1 | 0% | 5,522 | 5,056 | -8% | 0 | 0 | — |
case-07 | fail→fail | 22,812 | 17,175 | -25% | 1 | 1 | 0% | 3,344 | 1,233 | -63% | 0 | 0 | — |
case-08 | fail→fail | 18,414 | 28,367 | +54% | 1 | 1 | 0% | 2,601 | 3,503 | +35% | 0 | 0 | — |
case-09 | fail→fail | 18,722 | 22,777 | +22% | 1 | 1 | 0% | 2,753 | 2,453 | -11% | 0 | 0 | — |
case-10 | fail→fail | 21,078 | 22,329 | +6% | 1 | 1 | 0% | 2,977 | 3,329 | +12% | 0 | 0 | — |
case-11 | fail→fail | 18,469 | 35,782 | +94% | 1 | 1 | 0% | 2,818 | 4,700 | +67% | 0 | 0 | — |
case-12 | fail→fail | 19,748 | 25,102 | +27% | 1 | 1 | 0% | 2,923 | 3,879 | +33% | 0 | 0 | — |
case-13 | fail→fail | 21,579 | 23,781 | +10% | 1 | 1 | 0% | 3,188 | 3,777 | +18% | 0 | 0 | — |
case-14 | fail→fail | 25,683 | 17,527 | -32% | 1 | 1 | 0% | 3,961 | 2,893 | -27% | 0 | 0 | — |
case-15 | fail→fail | 17,481 | 18,584 | +6% | 1 | 1 | 0% | 2,635 | 2,802 | +6% | 0 | 0 | — |
case-16 | fail→fail | 21,740 | 27,250 | +25% | 1 | 1 | 0% | 3,441 | 3,737 | +9% | 0 | 0 | — |
case-17 | fail→fail | 20,652 | 14,846 | -28% | 1 | 1 | 0% | 3,106 | 2,369 | -24% | 0 | 0 | — |
case-18 | fail→fail | 22,092 | 23,421 | +6% | 1 | 1 | 0% | 3,326 | 2,476 | -26% | 0 | 0 | — |
case-19 | fail→fail | 17,973 | 21,302 | +19% | 1 | 1 | 0% | 2,863 | 3,336 | +17% | 0 | 0 | — |
case-20 | fail→fail | 21,195 | 21,541 | +2% | 1 | 1 | 0% | 3,202 | 3,420 | +7% | 0 | 0 | — |
case-21 | fail→fail | 21,448 | 17,418 | -19% | 1 | 1 | 0% | 3,237 | 2,651 | -18% | 0 | 0 | — |
case-22 | fail→fail | 18,060 | 26,671 | +48% | 1 | 1 | 0% | 2,586 | 3,944 | +53% | 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 0 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.