Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Identify orthogonal axes along which a method's validity might vary. Ensures axes are independent, measurable, and span the relevant parameter space.
.claude/skills/yogsoth-ai-variation-axis-definition/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✓→✗ | ▼ Worse | 17% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 71% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 4% | 0% |
| case-11 | ✗→✗ | ＝ Same ✗ | 11% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -9% | 0% |
Define the dimensions along which to test validity.
Subagent — spawned via subagent-spawning/spawn-agent.
One unit = one axis definition pass.
<!-- 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-04 | pass→pass | 20,057 | 29,400 | +47% | 1 | 1 | 0% | 2,863 | 4,907 | +71% | 0 | 0 | — |
case-11 | fail→fail | 21,958 | 24,680 | +12% | 1 | 1 | 0% | 2,741 | 3,032 | +11% | 0 | 0 | — |
case-05 | pass→fail | 20,163 | 29,622 | +47% | 1 | 1 | 0% | 3,871 | 4,536 | +17% | 0 | 0 | — |
case-01 | fail→fail | 20,347 | 18,229 | -10% | 1 | 1 | 0% | 2,541 | 2,322 | -9% | 0 | 0 | — |
case-02 | fail→fail | 26,177 | 25,167 | -4% | 1 | 1 | 0% | 3,374 | 3,297 | -2% | 0 | 0 | — |
case-03 | pass→pass | 52,925 | 26,906 | -49% | 1 | 1 | 0% | 4,293 | 4,471 | +4% | 0 | 0 | — |
case-06 | fail→fail | 23,653 | 26,112 | +10% | 1 | 1 | 0% | 2,938 | 3,745 | +27% | 0 | 0 | — |
case-07 | fail→fail | 19,859 | 22,396 | +13% | 1 | 1 | 0% | 2,373 | 3,067 | +29% | 0 | 0 | — |
case-08 | fail→fail | 28,184 | 29,453 | +5% | 1 | 1 | 0% | 3,875 | 4,464 | +15% | 0 | 0 | — |
case-09 | fail→fail | 25,416 | 31,497 | +24% | 1 | 1 | 0% | 3,212 | 4,252 | +32% | 0 | 0 | — |
case-10 | fail→fail | 26,260 | 26,609 | +1% | 1 | 1 | 0% | 3,348 | 3,582 | +7% | 0 | 0 | — |
case-12 | fail→fail | 22,749 | 36,391 | +60% | 1 | 1 | 0% | 2,868 | 2,441 | -15% | 0 | 0 | — |
case-13 | fail→fail | 21,110 | 18,418 | -13% | 1 | 1 | 0% | 2,627 | 2,185 | -17% | 0 | 0 | — |
case-14 | fail→fail | 17,737 | 19,186 | +8% | 1 | 1 | 0% | 2,113 | 2,310 | +9% | 0 | 0 | — |
case-15 | fail→fail | 26,558 | 22,043 | -17% | 1 | 1 | 0% | 3,516 | 3,016 | -14% | 0 | 0 | — |
case-16 | fail→fail | 22,050 | 24,015 | +9% | 1 | 1 | 0% | 2,718 | 3,232 | +19% | 0 | 0 | — |
case-17 | fail→fail | 24,451 | 20,309 | -17% | 1 | 1 | 0% | 2,877 | 2,332 | -19% | 0 | 0 | — |
case-18 | fail→fail | 22,735 | 26,310 | +16% | 1 | 1 | 0% | 2,834 | 3,348 | +18% | 0 | 0 | — |
case-19 | fail→fail | 26,748 | 30,847 | +15% | 1 | 1 | 0% | 3,388 | 4,284 | +26% | 0 | 0 | — |
case-20 | fail→fail | 23,819 | 18,883 | -21% | 1 | 1 | 0% | 2,897 | 2,142 | -26% | 0 | 0 | — |
case-21 | fail→fail | 22,953 | 28,839 | +26% | 1 | 1 | 0% | 2,849 | 3,680 | +29% | 0 | 0 | — |
case-22 | fail→fail | 24,641 | 33,644 | +37% | 1 | 1 | 0% | 2,953 | 5,024 | +70% | 0 | 0 | — |
case-23 | fail→fail | 22,802 | 22,820 | +0% | 1 | 1 | 0% | 2,601 | 2,857 | +10% | 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. 23 cases were attempted. The headline lift of -4 percentage points is the difference between those two pass rates over the 23 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.