Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when a campaign is live and needs a new asset the same day — creative fatigue, CTR decay, or testing a new angle: stripped-down input, duplicate the winning creative, one-page template, hard SLA. Trigger on 'quick brief', 'need an asset fast', 'creative fatigue', 'urgent creative', 'asset today', 'new creative fast', or 'swap the creative'.
.claude/skills/minhnv0807-48-quick-visual-brief-global/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 96% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 83% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 179% | 0% |
Use this when a campaign is live and an asset is needed urgently: speed beats perfection. A variant running in four hours beats a perfect asset delivered in two days. The brief needs to be enough to build, not enough to be beautiful.
| Situation | Use | |-----------|-----| | Creative fatigue, CTR decay, high frequency, variant needed today | This skill (48) | | Testing a new angle on a creative that already works | This skill (48) | | Entirely new asset, new concept | 42-image-brief-global (full brief) | | Landing page, video, email | 12, 44, 49 — these cannot be rushed | | No winning creative to build from | 42-image-brief-global |
Ask once, in a single message, up to 4 questions:
21-ads-audit-global.Which live creative has the highest CTR? That is the base. Get the link to the source file (Canva, Figma, Adobe Express) — always duplicate the source, never an exported file.
Use the template below. Fill in only the "specific changes" table — everything else inherits from the base.
Duplicate the base, change exactly the listed elements, export at the stated sizes, submit.
Right message? Right size? CTA present? If yes, push it live. No feedback within one hour means the designer treats it as approved and submits the final file.
File name: quick-brief-[campaign]-[YYYYMMDD].md
markdown# QUICK BRIEF — [Asset name] — [date] Deadline: [HH:MM on date] (not flexible) Build time: 2-4 hours Base: [link to source file — the highest-CTR creative] ## Base template - Duplicate from: [link] - Keep unchanged: [layout / background / logo / fonts] - Change only: [copy / color / image / CTA — pick one main change] ## Specific changes | Element | Current | New | |---------|---------|-----| | Headline | | | | Subtext | | | | CTA | | | | Visual / image | | | ## Sizes required - [ ] 1080x1080 — Meta Feed - [ ] 1080x1920 — Story / Reels (if needed) - [ ] 1200x628 — link banner (if needed) ## Export and submission - Format: JPG - Submit via: [Slack channel / shared Drive folder] - File name: QUICK_[Campaign]_[Date]_v1.jpg ## Review - Three checks only: right message / right size / CTA present - Feedback within: 30 minutes of submission - No feedback within one hour = approved ## Context (1-2 sentences) [CTR has decayed from X% / frequency has passed the threshold / testing a new angle: ...]
| Step | Owner | Time | |------|-------|------| | Write the quick brief | Lead / media buyer | 30 minutes from spotting the problem | | Build the asset | Designer | 2-4 hours | | Review | Lead / media buyer | 30 minutes after submission | | No feedback | — | After one hour, treat as approved | | Launch-day standby | Designer | Respond to requests within 2 hours |
| Signal | Typical threshold | Action | |--------|-------------------|--------| | CTR decay | Down 25-30% or more from the creative's own first-week baseline | New variant on the same offer, new hook | | Frequency climbing | Above roughly 2.5-3.0 in a 7-day window on a prospecting set | Refresh the creative, not the audience | | CPA drifting up | Materially above the campaign's own target, with stable CVR | Test a new angle before touching bids | | A new angle worth testing | Any time a hypothesis emerges from comments or support tickets | One variable changed, per 19-ab-test-setup-global |
Use the account's own trailing baseline as the reference, not an external benchmark — decay is relative to how the creative started.
21-ads-audit-global: the data source for creative fatigue (frequency, CTR) and for identifying the winner to clone.19-ab-test-setup-global: quick variants should still fit a clean test frame — one change equals one variable.41-campaign-asset-list-global: add the new asset row and update its status.42-image-brief-global: fallback when there is no winner to clone — write the full brief.50-asset-resize-global: once a variant wins, resize it into the remaining sizes.47-design-review-global: urgent assets get the 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→pass | 14,748 | 15,025 | +2% | 1 | 1 | 0% | 1,850 | 3,621 | +96% | 0 | 0 | — |
case-02 | fail→fail | 17,501 | 13,485 | -23% | 1 | 1 | 0% | 2,217 | 3,313 | +49% | 0 | 0 | — |
case-03 | fail→pass | 10,842 | 10,327 | -5% | 1 | 1 | 0% | 1,931 | 3,532 | +83% | 0 | 0 | — |
case-04 | fail→pass | 14,699 | 6,660 | -55% | 1 | 1 | 0% | 2,318 | 2,644 | +14% | 0 | 0 | — |
case-05 | fail→fail | 11,514 | 17,909 | +56% | 1 | 1 | 0% | 1,757 | 3,738 | +113% | 0 | 0 | — |
case-06 | fail→fail | 16,946 | 11,041 | -35% | 1 | 1 | 0% | 1,797 | 3,297 | +83% | 0 | 0 | — |
case-07 | fail→fail | 12,200 | 17,944 | +47% | 1 | 1 | 0% | 1,993 | 3,753 | +88% | 0 | 0 | — |
case-17 | fail→pass | 15,648 | 5,018 | -68% | 1 | 1 | 0% | 2,687 | 2,549 | -5% | 0 | 0 | — |
case-08 | fail→fail | 8,489 | 10,191 | +20% | 1 | 1 | 0% | 1,280 | 3,501 | +174% | 0 | 0 | — |
case-09 | fail→pass | 7,813 | 10,603 | +36% | 1 | 1 | 0% | 1,252 | 3,488 | +179% | 0 | 0 | — |
case-10 | pass→pass | 12,694 | 5,928 | -53% | 1 | 1 | 0% | 2,131 | 2,700 | +27% | 0 | 0 | — |
case-11 | fail→pass | 12,389 | 5,188 | -58% | 1 | 1 | 0% | 1,905 | 2,512 | +32% | 0 | 0 | — |
case-12 | fail→pass | 9,044 | 6,628 | -27% | 1 | 1 | 0% | 1,364 | 2,702 | +98% | 0 | 0 | — |
case-13 | pass→pass | 13,200 | 5,418 | -59% | 1 | 1 | 0% | 2,127 | 2,445 | +15% | 0 | 0 | — |
case-14 | fail→pass | 18,489 | 10,896 | -41% | 1 | 1 | 0% | 2,956 | 3,350 | +13% | 0 | 0 | — |
case-15 | fail→fail | 9,382 | 2,747 | -71% | 1 | 1 | 0% | 1,578 | 2,137 | +35% | 0 | 0 | — |
case-16 | fail→pass | 12,605 | 8,320 | -34% | 1 | 1 | 0% | 1,981 | 2,940 | +48% | 0 | 0 | — |
case-18 | fail→fail | 5,166 | 2,741 | -47% | 1 | 1 | 0% | 1,020 | 2,148 | +111% | 0 | 0 | — |
case-19 | pass→pass | 8,889 | 3,740 | -58% | 1 | 1 | 0% | 1,671 | 2,292 | +37% | 0 | 0 | — |
case-20 | pass→pass | 8,666 | 3,910 | -55% | 1 | 1 | 0% | 1,600 | 2,295 | +43% | 0 | 0 | — |
case-21 | fail→pass | 8,357 | 3,644 | -56% | 1 | 1 | 0% | 1,354 | 2,200 | +62% | 0 | 0 | — |
case-22 | fail→pass | 8,489 | 4,671 | -45% | 1 | 1 | 0% | 1,407 | 2,382 | +69% | 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 +50 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.