Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Direct visual and creative work for campaigns, photography, illustration, video, and branded experiences. Use this skill whenever the user wants to brief a photographer, direct illustrators, plan a creative campaign, develop visual concepts, write a creative direction document, or evaluate creative work for fit. Triggers on art direction, photo brief, photography brief, illustration brief, campaign concept, creative concept, visual direction, mood board, look and feel, visual treatment, video di
.claude/skills/rampstackco-art-direction/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 49% | 0% |
Direct creative work that extends the brand into specific deliverables. Photography, illustration, video, motion, campaigns, environmental design.
This skill assumes brand identity is approved (brand-identity complete). Art direction is about applying and extending it, not defining it.
creative-direction instead). This skill briefs specific creative deliverables; creative-direction produces the structured aesthetic brief that this skill consumes.brand-identity)design-standards)content-and-copy or landing-page-copy)design-system)A creative brief covers five layers. Each must be clear before the brief leaves your hands.
What this creative work is fundamentally about.
A weak premise produces work that's pretty but says nothing. Spend time here.
The visual treatment.
For photography:
For illustration:
For video / motion:
Production-level direction.
Specifications:
Constraints:
How this creative scales across distribution.
Most creative needs to live in multiple places. Plan the variants up front.
Common variant set for a campaign:
For each variant, note: how the composition adapts, what gets cropped or repositioned, what assets are required.
The quality bar.
Technical:
Creative:
Default output is a creative brief at creative-brief-[project].md, written against one of the brief templates in references/. The template you pick is the structure: use references/photo-shoot-brief.md for photography, references/illustration-brief.md for illustration, and references/creative-brief-template.md for anything else or for a mixed production. Each template owns its own section list, section order, and reference-image count. Do not write to the five-layer outline below and hope it maps; the templates carry sections the layers do not, licensing and usage rights among them, and those sections are load-bearing before production starts, not after it.
The five layers are the thinking. Every template covers all five, in its own order:
Plus a separate moodboard or visual reference doc with images, at the count the chosen template gives.
references/creative-brief-template.md - Generic art direction brief template covering any production type (photo, illustration, video, animation, mixed).references/photo-shoot-brief.md - Detailed brief template for photography commissions.references/illustration-brief.md - Brief template for illustration commissions.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→pass | 31,806 | 21,429 | -33% | 1 | 1 | 0% | 5,251 | 5,556 | +6% | 0 | 0 | — |
case-01 | fail→fail | 29,123 | 20,977 | -28% | 1 | 1 | 0% | 4,478 | 5,178 | +16% | 0 | 0 | — |
case-02 | fail→fail | 24,582 | 22,341 | -9% | 1 | 1 | 0% | 3,871 | 5,581 | +44% | 0 | 0 | — |
case-04 | fail→pass | 21,408 | 7,213 | -66% | 1 | 1 | 0% | 3,278 | 3,191 | -3% | 0 | 0 | — |
case-05 | fail→pass | 18,984 | 15,193 | -20% | 1 | 1 | 0% | 3,258 | 4,514 | +39% | 0 | 0 | — |
case-06 | fail→fail | 12,599 | 8,372 | -34% | 1 | 1 | 0% | 2,056 | 3,309 | +61% | 0 | 0 | — |
case-07 | fail→pass | 19,231 | 17,874 | -7% | 1 | 1 | 0% | 2,956 | 4,871 | +65% | 0 | 0 | — |
case-08 | fail→fail | 16,373 | 18,558 | +13% | 1 | 1 | 0% | 2,675 | 4,847 | +81% | 0 | 0 | — |
case-09 | fail→fail | 18,451 | 24,066 | +30% | 1 | 1 | 0% | 2,894 | 5,768 | +99% | 0 | 0 | — |
case-10 | pass→pass | 13,309 | 13,491 | +1% | 1 | 1 | 0% | 1,993 | 4,075 | +104% | 0 | 0 | — |
case-11 | fail→pass | 17,527 | 11,230 | -36% | 1 | 1 | 0% | 2,393 | 3,575 | +49% | 0 | 0 | — |
case-12 | fail→fail | 16,663 | 15,072 | -10% | 1 | 1 | 0% | 2,562 | 4,224 | +65% | 0 | 0 | — |
case-13 | pass→pass | 15,898 | 12,051 | -24% | 1 | 1 | 0% | 2,148 | 3,630 | +69% | 0 | 0 | — |
case-14 | pass→pass | 16,295 | 11,398 | -30% | 1 | 1 | 0% | 2,266 | 3,578 | +58% | 0 | 0 | — |
case-15 | pass→pass | 11,054 | 3,257 | -71% | 1 | 1 | 0% | 1,814 | 2,562 | +41% | 0 | 0 | — |
case-16 | fail→pass | 8,083 | 2,896 | -64% | 1 | 1 | 0% | 1,166 | 2,494 | +114% | 0 | 0 | — |
case-17 | pass→pass | 14,032 | 10,347 | -26% | 1 | 1 | 0% | 2,027 | 3,541 | +75% | 0 | 0 | — |
case-18 | pass→pass | 18,128 | 17,182 | -5% | 1 | 1 | 0% | 2,685 | 4,517 | +68% | 0 | 0 | — |
case-19 | pass→pass | 14,807 | 12,948 | -13% | 1 | 1 | 0% | 2,038 | 3,591 | +76% | 0 | 0 | — |
case-20 | pass→pass | 15,976 | 13,291 | -17% | 1 | 1 | 0% | 2,138 | 4,011 | +88% | 0 | 0 | — |
case-21 | pass→pass | 13,185 | 11,799 | -11% | 1 | 1 | 0% | 1,969 | 3,626 | +84% | 0 | 0 | — |
case-22 | fail→pass | 8,530 | 1,386 | -84% | 1 | 1 | 0% | 1,361 | 2,222 | +63% | 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. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 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.