Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Assumption Destruction Campaign — open new solution spaces by negating, reversing, and challenging fundamental assumptions.
.claude/skills/yogsoth-ai-assumption-destruction/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 99% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 208% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 123% | 0% |
Open new solution spaces by negating, reversing, and challenging fundamental assumptions.
| Strategy | Signal Keywords | |----------|----------------| | axiom-negation | negate, suspend, PO, de Bono, what if not, challenge axiom | | reverse-brainstorming | worse, reverse, anti-solution, how to fail, sabotage | | worst-method-inversion | worst possible, terrible solution, invert bad, flip failure | | anti-benchmark | best practice, industry standard, benchmark, conventional wisdom | | sacred-cow-hunting | unquestioned, sacred cow, taboo, never challenged, dogma |
| Strategy | Description | |----------|-------------| | axiom-negation | Identify and suspend fundamental assumptions via de Bono PO | | reverse-brainstorming | How to make it worse? → reverse for solutions | | worst-method-inversion | Design worst possible solution → extract insights → invert | | anti-benchmark | Challenge industry best practices' hidden assumptions | | sacred-cow-hunting | Find and challenge domain's unquestioned beliefs |
| Tactic | Description | |--------|-------------| | provocation-generation | Generate PO provocations and extract constructive movement (shared) | | assumption-enumeration | Surface, perturb, and prioritize assumptions by disruption potential | | inversion-protocol | Reverse statements → extract insights → build constructive alternatives |
| SOP | Description | |-----|-------------| | assumption-perturbation | Perturb each assumption, observe system response | | reversal-generation | Systematically reverse positive statements | | worst-case-design | Design the worst possible solution | | inversion-extraction | Extract constructive insights from worst solutions | | benchmark-challenge | Identify and negate benchmark assumptions | | sacred-cow-identification | Find domain's unquestioned beliefs | | constructive-rebellion | Build constructive alternatives from destructive negation | | destruction-synthesis | Synthesize all assumption destruction outputs |
| Strategy | web-search | web-research | paper-overview | paper-search | paper-research | |----------|-----------|-------------|---------------|-------------|---------------| | axiom-negation | 20 | 5 | 20 | 12 | 5 | | reverse-brainstorming | 15 | 5 | 15 | 10 | 3 | | worst-method-inversion | 15 | 5 | 15 | 10 | 3 | | anti-benchmark | 25 | 10 | 25 | 15 | 5 | | sacred-cow-hunting | 30 | 10 | 25 | 15 | 8 |
| Tool | Server | Purpose | |------|--------|---------| | brave_web_search | brave-search | General web search for methods and examples | | brave_llm_context | brave-search | Deep content extraction from web pages | | apify/rag-web-browser | apify | Full page scraping for detailed content | | get_paper_content | alphaxiv | Read academic paper content | | discover_papers | alphaxiv | Find relevant research papers | | relevanceSearch | semantic-scholar | Search academic literature | | paper | semantic-scholar | Get paper details | | citations | semantic-scholar | Trace citation networks |
| Tactic | Role | |--------|------| | provocation-generation | Generate PO provocations and extract movement (shared) | | assumption-enumeration | Surface and prioritize assumptions by disruption potential | | inversion-protocol | Reverse → extract → build constructive alternatives |
| SOP | Role | |-----|------| | assumption-perturbation | Perturb assumptions and observe system response | | reversal-generation | Systematically reverse positive statements | | worst-case-design | Design worst possible solution | | inversion-extraction | Extract constructive insights from inversions | | benchmark-challenge | Negate benchmark assumptions | | sacred-cow-identification | Find unquestioned beliefs | | constructive-rebellion | Build alternatives from negation | | destruction-synthesis | Final synthesis of all outputs |
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Strategy | When to use | | --- | --- | | anti-benchmark | Challenge industry best practices' hidden assumptions. Deconstruct benchmarks to reveal unexamined constraints. | | axiom-negation | Identify and suspend fundamental assumptions via de Bono PO. Systematically negate axioms to reveal hidden solution spaces. | | reverse-brainstorming | How to make it worse? → reverse for solutions. Generate anti-solutions then invert to discover novel approaches. | | sacred-cow-hunting | Find and challenge domain's unquestioned beliefs. Systematic identification and productive violation of dogma. | | worst-method-inversion | Design worst possible solution → extract insights → invert. Deliberate failure design as creative catalyst. |
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | creative-ideation-assumption-enumeration | Surface, perturb, and prioritize assumptions by disruption potential. Orchestrates assumption surfacing → perturbation → sacred cow identification → prioritization. | | creative-ideation-provocation-generation | Generate PO provocations and extract constructive movement. Orchestrates assumption surfacing → provocation creation → movement extraction → idea formation. | | evaluation-filtering | Multi-dimensional evaluation and tiered filtering of generated ideas. Orchestrates novelty assessment → feasibility check → ranking → selection. | | inversion-protocol | Reverse statements → extract insights → build constructive alternatives. Systematic inversion pipeline from negation to innovation. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | context-checkpoint | Append research process and results to the current Phase's context file. Covers both process and results with genuine substance. Use this skill at plan-designated checkpoint points — typically after each strategy completes or at key decision nodes within a research Phase. | | context-init | Create a new context file for a research Phase. Called once at Phase start to initialize the file that subsequent context-checkpoint calls will append to. Use this skill whenever a new research Phase begins and a fresh context file is needed. | | creative-ideation-assumption-surfacing | Enumerate implicit assumptions in a problem statement or existing solution. Produces categorized assumption inventory (physical, social, temporal, economic, technical). | | creative-ideation-novelty-scoring | Score ideas on novelty dimensions — structural distance from known solutions, conceptual surprise, domain-crossing depth. Produces ranked novelty assessment. | | creative-ideation-saturation-detection | Determine when additional ideation yields diminishing returns. Analyzes latest idea batch against existing corpus to judge continue/near-saturation/saturated. | | idea-synthesis | Synthesize diverse ideas into coherent solution concepts. Combines fragments from multiple ideation passes into structured, actionable ideas with clear mechanism descriptions. | | po-provocation | Generate PO (Provocative Operation) statements per de Bono's lateral thinking. Creates deliberately illogical provocations to escape dominant thinking patterns. |
<!-- 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→pass | 16,482 | 8,182 | -50% | 1 | 1 | 0% | 2,377 | 3,059 | +29% | 0 | 0 | — |
case-06 | fail→pass | 14,420 | 5,982 | -59% | 1 | 1 | 0% | 2,085 | 2,658 | +27% | 0 | 0 | — |
case-02 | fail→pass | 9,283 | 6,004 | -35% | 1 | 1 | 0% | 1,326 | 2,633 | +99% | 0 | 0 | — |
case-03 | pass→pass | 14,191 | 6,666 | -53% | 1 | 1 | 0% | 1,958 | 2,736 | +40% | 0 | 0 | — |
case-04 | pass→pass | 11,935 | 10,662 | -11% | 1 | 1 | 0% | 1,718 | 3,413 | +99% | 0 | 0 | — |
case-05 | pass→pass | 15,902 | 16,008 | +1% | 1 | 1 | 0% | 2,250 | 4,013 | +78% | 0 | 0 | — |
case-07 | pass→pass | 15,730 | 6,393 | -59% | 1 | 1 | 0% | 2,241 | 2,804 | +25% | 0 | 0 | — |
case-08 | pass→pass | 14,283 | 4,103 | -71% | 1 | 1 | 0% | 2,012 | 2,385 | +19% | 0 | 0 | — |
case-09 | pass→pass | 15,921 | 7,256 | -54% | 1 | 1 | 0% | 2,318 | 2,788 | +20% | 0 | 0 | — |
case-10 | pass→pass | 11,749 | 4,341 | -63% | 1 | 1 | 0% | 1,760 | 2,383 | +35% | 0 | 0 | — |
case-11 | pass→pass | 7,203 | 2,162 | -70% | 1 | 1 | 0% | 1,127 | 2,017 | +79% | 0 | 0 | — |
case-12 | pass→fail | 7,648 | 2,148 | -72% | 1 | 1 | 0% | 1,186 | 2,091 | +76% | 0 | 0 | — |
case-13 | pass→pass | 10,448 | 2,717 | -74% | 1 | 1 | 0% | 1,553 | 2,181 | +40% | 0 | 0 | — |
case-14 | pass→pass | 18,552 | 3,031 | -84% | 1 | 1 | 0% | 2,902 | 2,221 | -23% | 0 | 0 | — |
case-19 | fail→pass | 8,090 | 2,120 | -74% | 1 | 1 | 0% | 690 | 2,126 | +208% | 0 | 0 | — |
case-15 | pass→pass | 13,871 | 2,051 | -85% | 1 | 1 | 0% | 2,137 | 2,044 | -4% | 0 | 0 | — |
case-16 | pass→pass | 15,708 | 6,736 | -57% | 1 | 1 | 0% | 2,326 | 2,750 | +18% | 0 | 0 | — |
case-17 | pass→pass | 14,286 | 3,546 | -75% | 1 | 1 | 0% | 2,081 | 2,221 | +7% | 0 | 0 | — |
case-18 | pass→pass | 13,927 | 9,268 | -33% | 1 | 1 | 0% | 1,943 | 3,032 | +56% | 0 | 0 | — |
case-20 | fail→pass | 6,308 | 2,415 | -62% | 1 | 1 | 0% | 935 | 2,086 | +123% | 0 | 0 | — |
case-21 | fail→pass | 16,204 | 2,536 | -84% | 1 | 1 | 0% | 1,075 | 2,128 | +98% | 0 | 0 | — |
case-22 | pass→fail | 26,578 | 22,200 | -16% | 1 | 1 | 0% | 4,512 | 3,212 | -29% | 0 | 0 | — |
case-23 | pass→pass | 16,366 | 19,256 | +18% | 1 | 1 | 0% | 3,130 | 5,458 | +74% | 0 | 0 | — |
case-24 | pass→fail | 19,778 | 12,494 | -37% | 1 | 1 | 0% | 3,135 | 2,433 | -22% | 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. 24 cases were attempted, and 21 counted toward the lift figure. The other 3 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 +13 percentage points is the difference between those two pass rates over the 21 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.