Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build constructive alternatives from destructive negation. Transform violated assumptions into viable innovation directions.
.claude/skills/yogsoth-ai-constructive-rebellion/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-23 | ✓→✗ | ▼ Worse | -59% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 51% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 83% | 0% |
| case-04 | ✗→✗ | ＝ Same ✗ | 87% | 0% |
| case-05 | ✗→✗ | ＝ Same ✗ | 61% | 0% |
Build constructive alternatives from destructive negation.
Subagent — spawned via subagent-spawning/spawn-agent skill.
Constructive rebellion is the critical bridge from destruction to creation — it requires both creative generativity and disciplined feasibility thinking. Benefits from dedicated focus to avoid either premature dismissal or unconstrained fantasy.
<!-- 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 | fail→fail | 25,586 | 29,010 | +13% | 1 | 1 | 0% | 2,793 | 5,227 | +87% | 0 | 0 | — |
case-05 | fail→fail | 21,750 | 32,376 | +49% | 1 | 1 | 0% | 2,846 | 4,569 | +61% | 0 | 0 | — |
case-01 | fail→fail | 40,359 | 54,449 | +35% | 1 | 1 | 0% | 1,646 | 3,226 | +96% | 0 | 0 | — |
case-02 | fail→fail | 14,697 | 22,017 | +50% | 1 | 1 | 0% | 1,937 | 3,343 | +73% | 0 | 0 | — |
case-03 | fail→fail | 18,784 | 23,132 | +23% | 1 | 1 | 0% | 2,474 | 949 | -62% | 0 | 0 | — |
case-06 | fail→fail | 18,315 | 51,683 | +182% | 1 | 1 | 0% | 2,707 | 4,865 | +80% | 0 | 0 | — |
case-07 | fail→fail | 18,543 | 25,408 | +37% | 1 | 1 | 0% | 2,786 | 3,477 | +25% | 0 | 0 | — |
case-08 | fail→fail | 16,566 | 66,371 | +301% | 1 | 1 | 0% | 2,620 | 3,983 | +52% | 0 | 0 | — |
case-09 | fail→fail | 27,915 | 38,972 | +40% | 1 | 1 | 0% | 3,583 | 4,954 | +38% | 0 | 0 | — |
case-10 | fail→fail | 19,450 | 44,680 | +130% | 1 | 1 | 0% | 2,835 | 6,259 | +121% | 0 | 0 | — |
case-11 | fail→fail | 19,528 | 18,594 | -5% | 1 | 1 | 0% | 2,835 | 1,463 | -48% | 0 | 0 | — |
case-12 | fail→fail | 21,615 | 53,126 | +146% | 1 | 1 | 0% | 3,296 | 4,910 | +49% | 0 | 0 | — |
case-13 | fail→fail | 23,812 | 30,293 | +27% | 1 | 1 | 0% | 2,967 | 1,516 | -49% | 0 | 0 | — |
case-14 | fail→fail | 22,626 | 45,878 | +103% | 1 | 1 | 0% | 3,376 | 5,960 | +77% | 0 | 0 | — |
case-15 | fail→fail | 22,934 | 22,089 | -4% | 1 | 1 | 0% | 2,828 | 1,410 | -50% | 0 | 0 | — |
case-16 | fail→fail | 18,229 | 27,196 | +49% | 1 | 1 | 0% | 2,581 | 4,559 | +77% | 0 | 0 | — |
case-17 | fail→fail | 16,295 | 36,396 | +123% | 1 | 1 | 0% | 2,491 | 6,110 | +145% | 0 | 0 | — |
case-18 | fail→fail | 18,905 | 22,968 | +21% | 1 | 1 | 0% | 2,482 | 3,117 | +26% | 0 | 0 | — |
case-19 | fail→fail | 19,388 | 21,315 | +10% | 1 | 1 | 0% | 2,805 | 3,385 | +21% | 0 | 0 | — |
case-20 | fail→fail | 19,945 | 29,433 | +48% | 1 | 1 | 0% | 2,567 | 1,588 | -38% | 0 | 0 | — |
case-21 | pass→pass | 18,183 | 22,548 | +24% | 1 | 1 | 0% | 2,255 | 3,413 | +51% | 0 | 0 | — |
case-22 | pass→pass | 18,482 | 35,294 | +91% | 1 | 1 | 0% | 2,458 | 4,489 | +83% | 0 | 0 | — |
case-23 | pass→fail | 13,054 | 11,210 | -14% | 1 | 1 | 0% | 2,120 | 867 | -59% | 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 17 counted toward the lift figure. The other 6 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 -4 percentage points is the difference between those two pass rates over the 17 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.