Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →de Bono lateral escape sequence — identify dominant idea, generate provocations (escape/reversal/exaggeration/distortion), follow consequences to extract new framings. Breaks paradigm lock-in.
.claude/skills/yogsoth-ai-lateral-escape/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -40% | 0% |
Break paradigm lock-in through systematic provocation.
dominant-idea-identification → provocation-generation → consequence-following
Subagent: dominant-idea-identification, provocation-generation, consequence-following Import: web-research
First identify what everyone "knows" (the dominant idea), then systematically challenge it via PO provocations (4 types: escape, reversal, exaggeration, distortion), follow each provocation's consequences to see if viable new framings emerge.
The goal is not to prove the dominant idea wrong — it is to escape its gravitational pull and discover what becomes visible from outside it.
<HARD-GATE>
- Dominant ideas identified: >= 2
- Provocations generated: >= 4 (one per type minimum)
- Consequence chains followed: >= 3
- Viable new framings extracted: >= 1
</HARD-GATE><!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | consequence-following | Follow a provocation's logical consequences step by step to extract viable insights and new research directions. | | deep-insight-provocation-generation | Generate de Bono lateral thinking provocations to challenge dominant ideas using escape, reversal, exaggeration, and distortion. | | deep-insight-web-research | Deep web research with full page fetching via Apify. Import of web-browsing/web-research skill. Must fetch full page — no conclusions from previews. | | dominant-idea-identification | Identify dominant paradigms and assumptions constraining thinking in a research field. |
<!-- 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 | 45,640 | 73,319 | +61% | 1 | 1 | 0% | 5,640 | 7,139 | +27% | 0 | 0 | — |
case-02 | fail→pass | 51,649 | 50,641 | -2% | 1 | 1 | 0% | 4,868 | 6,707 | +38% | 0 | 0 | — |
case-03 | fail→pass | 32,761 | 49,082 | +50% | 1 | 1 | 0% | 4,190 | 6,416 | +53% | 0 | 0 | — |
case-04 | fail→fail | 42,777 | 42,848 | +0% | 1 | 1 | 0% | 6,927 | 6,701 | -3% | 0 | 0 | — |
case-05 | pass→pass | 42,109 | 41,538 | -1% | 1 | 1 | 0% | 6,655 | 5,992 | -10% | 0 | 0 | — |
case-06 | pass→pass | 30,301 | 35,339 | +17% | 1 | 1 | 0% | 3,654 | 4,728 | +29% | 0 | 0 | — |
case-07 | pass→pass | 24,253 | 36,058 | +49% | 1 | 1 | 0% | 3,065 | 6,553 | +114% | 0 | 0 | — |
case-08 | pass→fail | 19,852 | 63,591 | +220% | 1 | 1 | 0% | 3,193 | 945 | -70% | 0 | 0 | — |
case-09 | pass→pass | 48,518 | 62,009 | +28% | 1 | 1 | 0% | 7,062 | 8,623 | +22% | 0 | 0 | — |
case-10 | fail→pass | 25,901 | 23,600 | -9% | 1 | 1 | 0% | 3,448 | 4,173 | +21% | 0 | 0 | — |
case-11 | fail→pass | 48,882 | 29,507 | -40% | 1 | 1 | 0% | 8,237 | 4,906 | -40% | 0 | 0 | — |
case-12 | fail→pass | 41,798 | 26,701 | -36% | 1 | 1 | 0% | 7,130 | 4,944 | -31% | 0 | 0 | — |
case-13 | pass→pass | 24,383 | 42,530 | +74% | 1 | 1 | 0% | 3,551 | 7,557 | +113% | 0 | 0 | — |
case-14 | pass→pass | 25,329 | 53,196 | +110% | 1 | 1 | 0% | 3,818 | 4,831 | +27% | 0 | 0 | — |
case-15 | fail→fail | 49,223 | 9,848 | -80% | 1 | 1 | 0% | 8,237 | 817 | -90% | 0 | 0 | — |
case-16 | pass→pass | 42,523 | 54,162 | +27% | 1 | 1 | 0% | 7,114 | 6,405 | -10% | 0 | 0 | — |
case-17 | pass→pass | 19,226 | 26,778 | +39% | 1 | 1 | 0% | 3,008 | 4,439 | +48% | 0 | 0 | — |
case-18 | pass→pass | 25,870 | 17,419 | -33% | 1 | 1 | 0% | 4,284 | 2,679 | -37% | 0 | 0 | — |
case-19 | pass→pass | 28,429 | 55,273 | +94% | 1 | 1 | 0% | 4,190 | 8,620 | +106% | 0 | 0 | — |
case-20 | pass→fail | 43,027 | 34,665 | -19% | 1 | 1 | 0% | 8,226 | 1,564 | -81% | 0 | 0 | — |
case-21 | pass→pass | 18,346 | 42,106 | +130% | 1 | 1 | 0% | 4,479 | 8,622 | +92% | 0 | 0 | — |
case-22 | pass→fail | 11,836 | 26,793 | +126% | 1 | 1 | 0% | 2,058 | 4,930 | +140% | 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 19 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 +14 percentage points is the difference between those two pass rates over the 19 comparable cases. 4 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.