Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply CSH boundary critique — what is included/excluded, who benefits/is harmed, what expertise is privileged/marginalized. Identifies opportunities at the boundaries.
.claude/skills/yogsoth-ai-boundary-critique/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 109% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 85% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 111% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 118% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 74% | 0% |
Critically examine what a method's boundaries include and exclude.
| Base SOP | Target | ±10% Range | |----------|--------|------------| | web-search | 30 | 27–33 | | web-research | 15 | 13–17 | | paper-overview | 30 | 27–33 | | paper-search | 20 | 18–22 | | paper-research | 10 | 9–11 |
<HARD-GATE>
| SOP | Done | Target | % |
|-----|------|--------|---|
| web-search | ? | 30 | ? |
| web-research | ? | 15 | ? |
| paper-overview | ? | 30 | ? |
| paper-search | ? | 20 | ? |
| paper-research | ? | 10 | ? |
Budget Gate: OPEN/CLOSED (>=80% required to exit)
</HARD-GATE>Import: web-search, web-research, paper-overview, paper-search, paper-research Subagent: csh-12-question (reused from insight) Shared: assumption-surfacing
Apply CSH boundary critique to the method — what is included/excluded, who benefits/is harmed, what expertise is privileged/marginalized. Identify opportunities at the boundaries.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | csh-12-question | Apply Ulrich's Critical Systems Heuristics 12 questions across 4 dimensions (motivation, control, expertise, legitimacy) comparing is vs ought. | | deep-insight-assumption-surfacing | Systematically extract implicit assumptions from methods, frameworks, or arguments. Identifies what is taken for granted without explicit justification. |
<!-- 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 | 21,460 | 40,483 | +89% | 1 | 1 | 0% | 3,190 | 6,666 | +109% | 0 | 0 | — |
case-02 | fail→fail | 21,405 | 7,286 | -66% | 1 | 1 | 0% | 2,977 | 1,494 | -50% | 0 | 0 | — |
case-03 | fail→fail | 23,126 | 34,721 | +50% | 1 | 1 | 0% | 3,338 | 6,680 | +100% | 0 | 0 | — |
case-04 | pass→pass | 15,845 | 37,089 | +134% | 1 | 1 | 0% | 2,251 | 6,034 | +168% | 0 | 0 | — |
case-10 | fail→fail | 25,621 | 11,396 | -56% | 1 | 1 | 0% | 3,656 | 1,137 | -69% | 0 | 0 | — |
case-05 | pass→pass | 13,371 | 33,350 | +149% | 1 | 1 | 0% | 2,622 | 6,674 | +155% | 0 | 0 | — |
case-06 | pass→pass | 12,553 | 39,443 | +214% | 1 | 1 | 0% | 1,751 | 6,642 | +279% | 0 | 0 | — |
case-07 | fail→fail | 24,346 | 6,343 | -74% | 1 | 1 | 0% | 3,583 | 1,254 | -65% | 0 | 0 | — |
case-08 | fail→pass | 20,369 | 29,292 | +44% | 1 | 1 | 0% | 2,676 | 4,946 | +85% | 0 | 0 | — |
case-09 | fail→pass | 21,433 | 39,884 | +86% | 1 | 1 | 0% | 3,154 | 6,650 | +111% | 0 | 0 | — |
case-11 | fail→pass | 18,778 | 34,907 | +86% | 1 | 1 | 0% | 2,547 | 5,559 | +118% | 0 | 0 | — |
case-12 | fail→fail | 24,130 | 14,892 | -38% | 1 | 1 | 0% | 3,420 | 1,262 | -63% | 0 | 0 | — |
case-13 | fail→fail | 21,290 | 11,093 | -48% | 1 | 1 | 0% | 2,908 | 1,224 | -58% | 0 | 0 | — |
case-14 | fail→pass | 19,955 | 29,416 | +47% | 1 | 1 | 0% | 2,698 | 4,704 | +74% | 0 | 0 | — |
case-15 | fail→pass | 21,523 | 42,456 | +97% | 1 | 1 | 0% | 2,968 | 6,642 | +124% | 0 | 0 | — |
case-16 | fail→fail | 24,492 | 43,269 | +77% | 1 | 1 | 0% | 3,599 | 6,636 | +84% | 0 | 0 | — |
case-17 | fail→fail | 23,850 | 35,792 | +50% | 1 | 1 | 0% | 3,302 | 1,783 | -46% | 0 | 0 | — |
case-18 | fail→pass | 20,690 | 40,620 | +96% | 1 | 1 | 0% | 3,034 | 6,637 | +119% | 0 | 0 | — |
case-19 | fail→fail | 24,391 | 40,428 | +66% | 1 | 1 | 0% | 3,429 | 6,637 | +94% | 0 | 0 | — |
case-20 | fail→fail | 25,142 | 6,107 | -76% | 1 | 1 | 0% | 3,520 | 822 | -77% | 0 | 0 | — |
case-21 | fail→fail | 26,230 | 10,253 | -61% | 1 | 1 | 0% | 3,757 | 875 | -77% | 0 | 0 | — |
case-22 | fail→fail | 20,037 | 15,499 | -23% | 1 | 1 | 0% | 3,038 | 1,142 | -62% | 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 15 counted toward the lift figure. The other 7 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 +32 percentage points is the difference between those two pass rates over the 15 comparable cases.
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.