Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create visual moodboards from collected inspiration with iterative refinement. Use after trend research or website analysis to synthesize design direction before implementation.
.claude/skills/davepoon-moodboard-creator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 31% | 0% |
Create and refine visual moodboards that synthesize design inspiration into a cohesive direction.
Before jumping to code, create a moodboard that:
Collect inspiration from:
For each source, note:
From collected sources, extract:
Colors
Typography
UI Patterns
Mood/Atmosphere
Generate a structured moodboard:
markdown## Moodboard v1 - [Project Name] ### Inspiration Sources | Source | Key Takeaway | |--------|--------------| | [URL/Name 1] | [What we're taking from it] | | [URL/Name 2] | [What we're taking from it] | | [URL/Name 3] | [What we're taking from it] | ### Color Direction
Primary: #hex] - color name] Secondary: #hex] - color name] Accent: #hex] - color name] Background: #hex] - color name] Text: #hex] - color name]
### Typography Direction
- **Headlines**: [Font/style] - [weight, size notes]
- **Body**: [Font/style] - [readability notes]
- **Accents**: [Any special type treatments]
### UI Patterns to Incorporate
1. **[Pattern Name]**: [Description of how to use it]
2. **[Pattern Name]**: [Description of how to use it]
3. **[Pattern Name]**: [Description of how to use it]
### Layout Approach
- Grid system: [e.g., 12-column, bento, asymmetric]
- Spacing philosophy: [tight, airy, mixed]
- Section structure: [full-width, contained, alternating]
### Mood Keywords
[Keyword 1] | [Keyword 2] | [Keyword 3] | [Keyword 4]
### Visual References
[Descriptions of key screenshots/references]
### What to Avoid
- [Anti-pattern from inspiration that doesn't fit]
- [Style that would clash]Present moodboard to user and ask:
Based on feedback:
Continue until user approves.
When approved, create final moodboard summary:
markdown## FINAL Moodboard - [Project Name] ### Approved Direction [Summary of the design direction] ### Color Palette (Final) | Role | Hex | Usage | |------|-----|-------| | Primary | #xxx | Buttons, links, accents | | Secondary | #xxx | Hover states, icons | | Background | #xxx | Page background | | Surface | #xxx | Cards, modals | | Text Primary | #xxx | Headings, body | | Text Secondary | #xxx | Captions, muted | ### Typography (Final) - Headlines: [Font Name] - [weight] - Body: [Font Name] - [weight] - Monospace: [Font Name] (if needed) ### Key Patterns 1. [Pattern with implementation notes] 2. [Pattern with implementation notes] ### Ready for Implementation [Checkbox] Colors defined [Checkbox] Fonts selected [Checkbox] Layout approach set [Checkbox] User approved
If no visual sources available:
Final moodboard should directly inform:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 9,895 | 8,319 | -16% | 1 | 1 | 0% | 2,148 | 3,049 | +42% | 0 | 0 | — |
case-05 | pass→pass | 17,374 | 13,564 | -22% | 1 | 1 | 0% | 3,067 | 3,639 | +19% | 0 | 0 | — |
case-06 | pass→pass | 17,233 | 14,344 | -17% | 1 | 1 | 0% | 3,549 | 4,132 | +16% | 0 | 0 | — |
case-11 | pass→pass | 10,574 | 7,182 | -32% | 1 | 1 | 0% | 1,759 | 2,407 | +37% | 0 | 0 | — |
case-01 | fail→pass | 18,479 | 12,004 | -35% | 1 | 1 | 0% | 3,502 | 3,354 | -4% | 0 | 0 | — |
case-02 | fail→pass | 18,481 | 11,888 | -36% | 1 | 1 | 0% | 3,177 | 3,324 | +5% | 0 | 0 | — |
case-03 | fail→pass | 15,552 | 9,764 | -37% | 1 | 1 | 0% | 2,610 | 2,937 | +13% | 0 | 0 | — |
case-07 | fail→pass | 15,208 | 9,176 | -40% | 1 | 1 | 0% | 2,520 | 2,779 | +10% | 0 | 0 | — |
case-08 | fail→pass | 10,031 | 6,644 | -34% | 1 | 1 | 0% | 1,723 | 2,256 | +31% | 0 | 0 | — |
case-09 | fail→pass | 13,232 | 7,223 | -45% | 1 | 1 | 0% | 2,232 | 2,384 | +7% | 0 | 0 | — |
case-10 | pass→fail | 15,757 | 7,985 | -49% | 1 | 1 | 0% | 2,480 | 2,459 | -1% | 0 | 0 | — |
case-12 | fail→pass | 9,487 | 4,510 | -52% | 1 | 1 | 0% | 1,510 | 1,848 | +22% | 0 | 0 | — |
case-13 | pass→pass | 16,613 | 7,768 | -53% | 1 | 1 | 0% | 2,700 | 2,394 | -11% | 0 | 0 | — |
case-14 | pass→pass | 8,084 | 3,124 | -61% | 1 | 1 | 0% | 1,347 | 1,668 | +24% | 0 | 0 | — |
case-15 | fail→fail | 4,631 | 2,784 | -40% | 1 | 1 | 0% | 782 | 1,599 | +104% | 0 | 0 | — |
case-16 | pass→pass | 12,448 | 8,963 | -28% | 1 | 1 | 0% | 2,008 | 2,624 | +31% | 0 | 0 | — |
case-17 | fail→pass | 10,081 | 3,128 | -69% | 1 | 1 | 0% | 1,585 | 1,696 | +7% | 0 | 0 | — |
case-18 | fail→pass | 10,128 | 8,483 | -16% | 1 | 1 | 0% | 1,675 | 2,578 | +54% | 0 | 0 | — |
case-19 | fail→fail | 3,784 | 3,338 | -12% | 1 | 1 | 0% | 653 | 1,693 | +159% | 0 | 0 | — |
case-20 | fail→pass | 10,344 | 3,928 | -62% | 1 | 1 | 0% | 1,904 | 1,879 | -1% | 0 | 0 | — |
case-21 | pass→pass | 13,901 | 4,392 | -68% | 1 | 1 | 0% | 2,295 | 1,897 | -17% | 0 | 0 | — |
case-22 | fail→pass | 5,563 | 1,419 | -74% | 1 | 1 | 0% | 991 | 1,380 | +39% | 0 | 0 | — |
case-23 | pass→pass | 13,073 | 7,604 | -42% | 1 | 1 | 0% | 2,125 | 2,435 | +15% | 0 | 0 | — |
case-24 | pass→pass | 11,883 | 10,495 | -12% | 1 | 1 | 0% | 1,935 | 2,969 | +53% | 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. The headline lift of +42 percentage points is the difference between those two pass rates over the 24 comparable cases. 1 case got worse with the skill loaded, and it is 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.