Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Novice perspective — question the 'obvious' by adopting deliberate ignorance to reveal hidden complexity.
.claude/skills/yogsoth-ai-novice-perspective/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-21 | ✓→✗ | ▼ Worse | -57% | 0% |
| case-23 | ✓→✗ | ▼ Worse | -55% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 69% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | 62% | 0% |
Novice perspective: question the 'obvious'.
Subagent — spawned via subagent-spawning/spawn-agent skill.
Deliberate novice thinking requires suppressing expert knowledge, which benefits from a clean context without accumulated domain assumptions.
<!-- 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-02 | fail→fail | 36,384 | 12,311 | -66% | 1 | 1 | 0% | 1,213 | 1,965 | +62% | 0 | 0 | — |
case-01 | fail→fail | 20,478 | 37,507 | +83% | 1 | 1 | 0% | 2,461 | 1,725 | -30% | 0 | 0 | — |
case-03 | fail→fail | 11,565 | 34,130 | +195% | 1 | 1 | 0% | 978 | 1,757 | +80% | 0 | 0 | — |
case-04 | fail→fail | 40,144 | 35,222 | -12% | 1 | 1 | 0% | 3,077 | 4,959 | +61% | 0 | 0 | — |
case-05 | fail→fail | 19,093 | 51,655 | +171% | 1 | 1 | 0% | 1,172 | 6,521 | +456% | 0 | 0 | — |
case-06 | fail→fail | 37,973 | 62,352 | +64% | 1 | 1 | 0% | 2,775 | 1,135 | -59% | 0 | 0 | — |
case-07 | fail→fail | 32,548 | 39,129 | +20% | 1 | 1 | 0% | 1,622 | 538 | -67% | 0 | 0 | — |
case-08 | fail→fail | 86,003 | 154,592 | +80% | 1 | 1 | 0% | 1,571 | 1,107 | -30% | 0 | 0 | — |
case-09 | fail→fail | 38,613 | 32,968 | -15% | 1 | 1 | 0% | 2,577 | 516 | -80% | 0 | 0 | — |
case-10 | fail→fail | 13,454 | 70,324 | +423% | 1 | 1 | 0% | 1,340 | 642 | -52% | 0 | 0 | — |
case-11 | fail→fail | 16,527 | 26,868 | +63% | 1 | 1 | 0% | 1,873 | 1,048 | -44% | 0 | 0 | — |
case-12 | fail→fail | 13,694 | 16,574 | +21% | 1 | 1 | 0% | 1,240 | 543 | -56% | 0 | 0 | — |
case-13 | fail→fail | 13,431 | 12,840 | -4% | 1 | 1 | 0% | 1,243 | 441 | -65% | 0 | 0 | — |
case-14 | fail→fail | 13,290 | 24,752 | +86% | 1 | 1 | 0% | 817 | 1,033 | +26% | 0 | 0 | — |
case-15 | fail→fail | 16,590 | 21,836 | +32% | 1 | 1 | 0% | 1,607 | 861 | -46% | 0 | 0 | — |
case-16 | fail→fail | 21,453 | 14,636 | -32% | 1 | 1 | 0% | 2,470 | 569 | -77% | 0 | 0 | — |
case-17 | fail→fail | 19,361 | 19,541 | +1% | 1 | 1 | 0% | 2,132 | 470 | -78% | 0 | 0 | — |
case-18 | fail→fail | 14,134 | 20,453 | +45% | 1 | 1 | 0% | 1,330 | 1,199 | -10% | 0 | 0 | — |
case-19 | fail→fail | 8,822 | 23,384 | +165% | 1 | 1 | 0% | 1,278 | 1,149 | -10% | 0 | 0 | — |
case-20 | pass→pass | 29,758 | 48,197 | +62% | 1 | 1 | 0% | 3,795 | 6,402 | +69% | 0 | 0 | — |
case-21 | pass→fail | 20,336 | 25,003 | +23% | 1 | 1 | 0% | 2,460 | 1,053 | -57% | 0 | 0 | — |
case-22 | pass→pass | 40,953 | 49,333 | +20% | 1 | 1 | 0% | 6,584 | 8,162 | +24% | 0 | 0 | — |
case-23 | pass→fail | 17,866 | 26,667 | +49% | 1 | 1 | 0% | 2,231 | 994 | -55% | 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, and 5 counted toward the lift figure. The other 18 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 -9 percentage points is the difference between those two pass rates over the 5 comparable cases. 5 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.