Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Theorist perspective — assess theoretical foundations, formal rigor, and formalization opportunities.
.claude/skills/yogsoth-ai-theorist-hat/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✓→✗ | ▼ Worse | -65% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 35% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 39% | 0% |
| case-13 | ✗→✗ | ＝ Same ✗ | -81% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 12% | 0% |
Theorist perspective: assess theoretical foundations.
Subagent — spawned via subagent-spawning/spawn-agent skill.
Theoretical analysis requires deep formal reasoning and literature awareness that benefits from dedicated context.
<!-- 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-13 | fail→fail | 35,854 | 24,004 | -33% | 1 | 1 | 0% | 5,170 | 1,006 | -81% | 0 | 0 | — |
case-01 | fail→fail | 13,701 | 13,452 | -2% | 1 | 1 | 0% | 1,298 | 1,449 | +12% | 0 | 0 | — |
case-02 | fail→fail | 22,925 | 20,762 | -9% | 1 | 1 | 0% | 2,730 | 889 | -67% | 0 | 0 | — |
case-03 | fail→fail | 14,727 | 52,499 | +256% | 1 | 1 | 0% | 1,376 | 7,487 | +444% | 0 | 0 | — |
case-04 | fail→fail | 29,931 | 23,372 | -22% | 1 | 1 | 0% | 4,336 | 1,128 | -74% | 0 | 0 | — |
case-05 | fail→fail | 20,042 | 50,236 | +151% | 1 | 1 | 0% | 2,094 | 5,087 | +143% | 0 | 0 | — |
case-06 | fail→fail | 32,923 | 43,388 | +32% | 1 | 1 | 0% | 4,335 | 6,653 | +53% | 0 | 0 | — |
case-07 | fail→fail | 56,362 | 57,530 | +2% | 1 | 1 | 0% | 8,223 | 8,363 | +2% | 0 | 0 | — |
case-08 | fail→fail | 33,687 | 24,251 | -28% | 1 | 1 | 0% | 4,938 | 927 | -81% | 0 | 0 | — |
case-09 | fail→fail | 36,957 | 55,121 | +49% | 1 | 1 | 0% | 5,136 | 8,765 | +71% | 0 | 0 | — |
case-10 | fail→fail | 26,916 | 20,564 | -24% | 1 | 1 | 0% | 3,419 | 726 | -79% | 0 | 0 | — |
case-11 | fail→fail | 34,384 | 42,436 | +23% | 1 | 1 | 0% | 4,909 | 6,297 | +28% | 0 | 0 | — |
case-12 | fail→fail | 36,254 | 45,291 | +25% | 1 | 1 | 0% | 6,078 | 8,357 | +37% | 0 | 0 | — |
case-14 | fail→fail | 44,352 | 48,228 | +9% | 1 | 1 | 0% | 7,318 | 7,920 | +8% | 0 | 0 | — |
case-15 | fail→fail | 33,065 | 63,652 | +93% | 1 | 1 | 0% | 4,823 | 8,155 | +69% | 0 | 0 | — |
case-16 | fail→fail | 36,254 | 16,330 | -55% | 1 | 1 | 0% | 4,705 | 639 | -86% | 0 | 0 | — |
case-17 | fail→fail | 19,024 | 33,289 | +75% | 1 | 1 | 0% | 2,178 | 539 | -75% | 0 | 0 | — |
case-18 | fail→fail | 28,953 | 15,336 | -47% | 1 | 1 | 0% | 3,777 | 414 | -89% | 0 | 0 | — |
case-19 | fail→fail | 29,124 | 30,921 | +6% | 1 | 1 | 0% | 5,075 | 5,741 | +13% | 0 | 0 | — |
case-20 | pass→fail | 7,706 | 10,861 | +41% | 1 | 1 | 0% | 1,291 | 451 | -65% | 0 | 0 | — |
case-21 | pass→pass | 16,362 | 22,402 | +37% | 1 | 1 | 0% | 3,410 | 4,600 | +35% | 0 | 0 | — |
case-22 | pass→pass | 6,336 | 8,269 | +31% | 1 | 1 | 0% | 1,335 | 1,850 | +39% | 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 13 counted toward the lift figure. The other 9 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 -5 percentage points is the difference between those two pass rates over the 13 comparable cases. 6 cases got worse with the skill loaded, and they are included in that figure.
Other measured skills in the registry, with their headline benchmark lift.