Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when the user wants a design reviewed against a scoring rubric across 4 groups — brand consistency, hierarchy and readability and accessibility, conversion elements, platform fit — 10 points each. Output is specific per-issue feedback plus a verdict of approve, minor fix, or redo. Trigger on 'design review', 'review this design', 'design feedback', 'score this design', 'brand check', 'is this design good', or 'approve this design'.
.claude/skills/minhnv0807-47-design-review-global/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 67% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 107% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 82% | 0% |
Reviewing design is not about liking it. It is checking whether the design does its job. Beautiful work that does not convert has failed. Converting work that breaks the brand has also failed. Feedback must be specific enough for the designer to act without asking.
Read the brand guideline (46-brand-guideline-global) and the originating brief. If information is missing, ask up to 4 questions:
46-brand-guideline-global first.View it as a real user would — do not zoom into detail yet. Answer two questions: within three seconds, is it clear what this design is saying? Where does the eye land first?
Each group is worth 10 points, 40 total. Score against the checklist below; deduct in proportion to the impact of each failure.
Every issue is written as Issue] -> Why it is wrong] -> Exactly how to fix it].
State the verdict, list fixes in priority order, and name the condition for approval.
markdown# Design Review — [Asset name] — [date] ## Quick verdict | Criterion | Assessment | Note | |-----------|-----------|------| | On brand? | OK / Warning / Fail | | | Clear hierarchy? | OK / Warning / Fail | | | Readable on mobile? | OK / Warning / Fail | | | Meets the objective? | OK / Warning / Fail | | ## Total: [X/40] | Group | Score | Assessment | |-------|-------|-----------| | A. Brand consistency | /10 | | | B. Hierarchy, readability, accessibility | /10 | | | C. Conversion elements | /10 | | | D. Platform fit | /10 | | Verdict: APPROVE / MINOR FIX (fix per notes, no re-review needed) / REDO [section] (reason) ## A. Brand consistency [/10] - [ ] Logo uses the right version, position, and clear space - [ ] Colors match the brand palette exactly (not a near-match with the wrong value) - [ ] Fonts are [heading font] / [body font] - [ ] Visual tone matches the brand personality - [ ] No element violates a guideline do or don't - [ ] No generic stock imagery diluting the brand Issue: [issue] -> [why] -> [how to fix] ## B. Hierarchy, readability, accessibility [/10] - [ ] One clear focal point — no competing centers of attention - [ ] Logical reading order: headline -> subhead -> body -> CTA - [ ] The most important element carries enough visual weight (size, color, contrast) - [ ] The headline is readable in three seconds - [ ] Body type is at least 14px and readable on a phone - [ ] Text contrast meets WCAG AA: 4.5:1 for body, 3:1 for large text (24px+, or 18.66px+ bold) - [ ] Icons and meaningful graphics meet 3:1 contrast; meaning is never carried by color alone - [ ] Alt text is drafted and ready for the post or email; video has captions - [ ] Not overloaded — three messages maximum, enough negative space - [ ] Spacing is consistent (8px grid); no typos, broken glyphs, or missing characters Issue: [issue] -> [why] -> [how to fix] ## C. Conversion elements [/10] - [ ] CTA stands out, contrasts with its ground, uses a verb plus a benefit - [ ] USP or offer is explicit (at BOFU the viewer knows exactly what they get) - [ ] Trust elements present where needed: figures, testimonial, badge - [ ] Urgency or scarcity present if the campaign requires it, and genuinely true - [ ] The hook is strong enough (at TOFU, it stops the scroll in about 1.5 seconds) - [ ] Claims are substantiated; disclosures present where required Issue: [issue] -> [why] -> [how to fix] ## D. Platform fit [/10] - [ ] Dimensions match the channel spec - [ ] Safe zones respected — text and CTA are not covered by UI or cropped - [ ] Mobile-friendly (readable at 375px width) - [ ] File size and format correct (JPG/PNG/MP4/PDF) - [ ] Overlay text kept light enough not to crowd the frame - [ ] Email: 600px width, CTA button at least 44px tall Issue: [issue] -> [why] -> [how to fix] ## Action summary Keep: [element] — [why it works] Fix now (must fix): 1. [issue] -> [fix] -> [deadline] Consider (should fix): 1. [suggestion] -> [reason] Approve when: [specific condition]
| Channel | Dimensions | Safe zone | |---------|-----------|-----------| | Meta Feed (FB/IG) | 1080x1080 (1:1) / 1080x1350 (4:5) | 20px each edge | | Meta Story / Reels | 1080x1920 (9:16) | 250px top / 250px bottom | | TikTok / Reels | 1080x1920 (9:16) | 150px top / 300px bottom, ~140px right | | LinkedIn | 1200x627 (1.91:1) | 40px each edge | | Pinterest | 1000x1500 (2:3) | 60px each edge | | YouTube thumbnail | 1280x720 (16:9) | avoid the bottom-right timestamp overlay | | Email header | 600x200 | 10px each edge | | Landing page hero | 1440px desktop / 375px mobile | headline and CTA visible without scrolling |
42, 43, 45) and rewriting it before rebuilding.48-quick-visual-brief-global applies (30 minutes).46-brand-guideline-global: the basis for group A — read it before reviewing.42-image-brief-global, 43-carousel-brief-global, 45-design-tool-brief-global, 12-landing-page-brief-global: the originating brief to check the design against.49-html-email-template-global: email-specific rendering and client checks before sending.41-campaign-asset-list-global: after approval, update the asset status (REVIEW to APPROVED).48-quick-visual-brief-global: urgent in-flight assets get a 3-point check, not the full 40-point review.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 22,325 | 29,067 | +30% | 1 | 1 | 0% | 3,639 | 6,406 | +76% | 0 | 0 | — |
case-02 | fail→pass | 30,558 | 30,451 | -0% | 1 | 1 | 0% | 4,756 | 6,367 | +34% | 0 | 0 | — |
case-03 | pass→pass | 30,817 | 80,757 | +162% | 1 | 1 | 0% | 3,464 | 6,292 | +82% | 0 | 0 | — |
case-04 | pass→pass | 35,379 | 51,725 | +46% | 1 | 1 | 0% | 3,319 | 6,924 | +109% | 0 | 0 | — |
case-05 | pass→pass | 17,965 | 11,429 | -36% | 1 | 1 | 0% | 2,281 | 4,224 | +85% | 0 | 0 | — |
case-06 | fail→pass | 12,515 | 23,203 | +85% | 1 | 1 | 0% | 1,932 | 3,222 | +67% | 0 | 0 | — |
case-07 | pass→pass | 10,867 | 9,984 | -8% | 1 | 1 | 0% | 1,475 | 3,164 | +115% | 0 | 0 | — |
case-12 | fail→pass | 14,573 | 12,887 | -12% | 1 | 1 | 0% | 1,656 | 3,426 | +107% | 0 | 0 | — |
case-08 | pass→pass | 14,019 | 7,223 | -48% | 1 | 1 | 0% | 1,978 | 3,238 | +64% | 0 | 0 | — |
case-09 | pass→pass | 14,627 | 16,711 | +14% | 1 | 1 | 0% | 1,469 | 3,127 | +113% | 0 | 0 | — |
case-10 | pass→pass | 11,988 | 7,463 | -38% | 1 | 1 | 0% | 1,240 | 3,153 | +154% | 0 | 0 | — |
case-11 | pass→pass | 8,537 | 6,305 | -26% | 1 | 1 | 0% | 1,356 | 3,141 | +132% | 0 | 0 | — |
case-13 | pass→pass | 12,093 | 5,995 | -50% | 1 | 1 | 0% | 1,996 | 3,074 | +54% | 0 | 0 | — |
case-14 | fail→pass | 16,187 | 11,621 | -28% | 1 | 1 | 0% | 1,705 | 3,194 | +87% | 0 | 0 | — |
case-15 | pass→pass | 8,404 | 2,829 | -66% | 1 | 1 | 0% | 1,222 | 2,654 | +117% | 0 | 0 | — |
case-16 | pass→pass | 8,491 | 7,409 | -13% | 1 | 1 | 0% | 1,288 | 2,500 | +94% | 0 | 0 | — |
case-17 | pass→pass | 10,576 | 12,421 | +17% | 1 | 1 | 0% | 1,646 | 3,271 | +99% | 0 | 0 | — |
case-18 | pass→pass | 11,200 | 8,766 | -22% | 1 | 1 | 0% | 1,632 | 2,714 | +66% | 0 | 0 | — |
case-19 | pass→pass | 9,991 | 14,231 | +42% | 1 | 1 | 0% | 1,725 | 3,581 | +108% | 0 | 0 | — |
case-20 | fail→fail | 12,306 | 3,115 | -75% | 1 | 1 | 0% | 1,021 | 2,652 | +160% | 0 | 0 | — |
case-21 | pass→pass | 15,110 | 3,842 | -75% | 1 | 1 | 0% | 1,652 | 2,706 | +64% | 0 | 0 | — |
case-22 | pass→pass | 12,928 | 8,679 | -33% | 1 | 1 | 0% | 1,741 | 2,671 | +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. 22 cases were attempted. The headline lift of +18 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.