メニュー
Apply universal design principles to create or improve designs. Domain-adaptive, generative, grounded in exemplars. Embeds generate → critique → iterate.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Route any input through a branching question tree to narrow down the optimal response strategy before writing. Two stages — PERCEIVE (classify input) then ACT (select response). Covers all prompt types.
Generate exhaustive guesses about user input using ALL search methods with coverage tracking. Guessing is SEARCH through possibility space. Tracks space created vs space covered to ensure comprehensive exploration.
Systematically evaluate and select from a set of guesses, options, or possibilities. Combines ARAW analysis with prioritization to determine which guesses are strong, weak, actionable, or eliminable.
Extract hidden assumptions from any content. Surfaces what must be true for claims to hold, enabling deeper analysis.
Sub-orchestrator for analytical requests. Routes to decomposition, systems analysis, comparison, risk assessment, or synthesis based on what kind of analysis is needed.
Assume Right - Deep recursive rightness search. For every claim, assume it's right — find what must follow, then assume THOSE implications are right too. Recurse until bedrock. Track every claim found.
| name | dsn - Design - Principled Design Creation and Improvement |
| description | Apply universal design principles to create or improve designs. Domain-adaptive, generative, grounded in exemplars. Embeds generate → critique → iterate. |
| output | {"format":"prose"} |
Input: $ARGUMENTS
Design is constrained optimization of arrangement. Given elements, constraints, and purpose — find the arrangement that best serves that purpose. This applies to visual, physical, digital, informational, and experiential design.
Purpose before aesthetics. Every element must serve the purpose. If removing an element doesn't degrade function or signal, remove it. This is parsimony — not minimalism (a Baroque church can be parsimonious if every ornament serves awe).
Exemplars before analysis. Before designing or critiquing, identify what good looks like. Exemplars provide a perceptual anchor that prevents analysis from drifting into "analytically sound but actually bad" territory.
Generate then evaluate. Produce candidate designs before analyzing them. Labels ("make it cleaner") are less useful than artifacts (actual CSS with specific values).
Trust the impression. When the design "feels wrong" but the individual elements check out, the impression is usually right. Investigate the relationships between elements, not the elements themselves.
Certainty gradient. Design decisions range from universal neuroscience (contrast creates hierarchy) to cultural convention (logo top-left) to arbitrary choice (exact shade of grey). Invest effort from the certain end. Don't agonize over arbitrary decisions — use convention.
Constraints improve quality. They reduce the search space. Embrace them.
1. UNDERSTAND → What's the purpose? Who's the audience? What are the constraints?
2. GROUND → What does good look like? (exemplars)
3. ASSESS → Where are we now? (if improving existing design)
4. PRIORITIZE → What to work on first? (domain effort ordering)
5. GENERATE → Produce candidate designs (/gen)
6. CRITIQUE → Evaluate candidates (/cri)
7. ITERATE → Improve best candidate based on critique
8. VERIFY → Does the final design satisfy purpose?
Steps 5-7 may loop until satisfied.
Always apply. Every domain.
| # | Principle | Test |
|---|---|---|
| 1 | PARSIMONY | Remove each element mentally. Does the design get worse? If not, remove it. |
| 2 | HIERARCHY | Can you tell what's most important within 3 seconds? |
| 3 | GROUPING | Are related things close and unrelated things separated? |
| 4 | PURPOSE ALIGNMENT | Does every element serve the goal? |
| 5 | COGNITIVE COST | How much mental effort does this require? Can it be less? |
| 6 | RHYTHM | Is spacing/proportion consistent and mathematical? |
| 7 | COMPOSITION | Is the whole more than the sum of parts? (Gestalt check) |
| 8 | CONSTRAINT GRATITUDE | Are constraints being used, not fought? |
These are default orderings. If your specific design has an obvious bottleneck that isn't at the top of this list, start there. The ordering is a heuristic, not a procedure.
Work from top to bottom. Top items produce the most quality per unit of effort.
For any design decision: "If I changed this, would the user notice and care?"
Convention defaults: standard grid, system fonts or proven web fonts, 4/8px spacing base, established color palette, platform-native interactions.
Every design outputs two things:
Signal is produced by execution quality. Pixel-perfect alignment, smooth animation, consistent spacing — these signal "we care about details." You can't fake signal cheaply (that's why it works — costly signaling).
| Context | Utility Weight | Signal Weight |
|---|---|---|
| Developer tool | 80% | 20% |
| Consumer app | 60% | 40% |
| Luxury/brand | 20% | 80% |
| Internal tool | 90% | 10% |
When improving an existing design:
For those who want precision:
Design D = (Elements E, Relationships R, Hierarchy H, Spatial arrangement S, Visual properties V, Purpose P)
Quality constraints:
1. Parsimony: ∀e ∈ E: utility(D) > utility(D \ {e})
2. Hierarchy: visual_difference(level_n, level_n+1) > perceptual_threshold
3. Grouping: related elements closer than unrelated
4. Rhythm: spacing values have low entropy (mathematical system)
5. Convention: deviation_cost < communication_gain
Objective function:
utility(D) = comprehension_speed × task_completion × signal_value − cognitive_cost − production_cost