motiful/product-shots / skills/product-shots-ad-creative
product-shots-ad-creative
VerifiedDesigns high-performing ad creatives across Instagram, Facebook, TikTok, LinkedIn, Google (Display / Demand Gen), YouTube, Pinterest, and X/Twitter — locking platform dimensions, safe zones, text overlay policy, character limits, industry visual DNA, ad objective hierarchy, composition patterns, CTA strategy, and prompt sanitation. Use when the user says "create an ad", "design an ad creative", "make an ad", "广告创意", "帮我做个广告图", "Instagram ad", "TikTok ad", "Facebook ad", "Google Display ad", "LinkedIn ad", "YouTube ad", "Pinterest ad", "promotional image", "campaign creative", "ad banner", "投放图", or any platform-specific ad request. Routed from `product-shots` when `asset_type == "ad"` or `is_promotion == True`.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 17.0 KB |
| references/ad-objective-rules.md | markdown | 5.2 KB |
| references/banned-words.md | markdown | 6.4 KB |
| references/color-and-cta-strategy.md | markdown | 8.4 KB |
| references/composition-patterns.md | markdown | 6.4 KB |
| references/failures-defaults-output.md | markdown | 7.3 KB |
| references/hard-constraints.md | markdown | 13.8 KB |
| references/industry-style-rules.md | markdown | 7.1 KB |
| references/multi-platform-output.md | markdown | 6.0 KB |
| references/platform-style-profiles.md | markdown | 7.5 KB |
| references/quick-start-and-fidelity.md | markdown | 6.2 KB |
| references/workflow-and-self-check.md | markdown | 11.4 KB |
references/workflow-and-self-check.md
View on GitHub---
name: workflow-and-self-check
description: Section 6 (Core Workflow — Six-Step Process). The procedural backbone of ad-creative — Step 1 Platform+Objective, Step 2 Industry+Brand, Step 3 Visual Asset Audit, Step 4 Composition+Color, Step 5 Platform Adaptation Layer, Step 6 Self-Check Gate (Quality Check Before Delivery, the 11-item mandatory checklist).
---
# Core Workflow — Six-Step Process
The procedural backbone of `product-shots-ad-creative`. The workflow is **self-contained** — it runs entirely inside this skill. The only external dependency is conditional pre-fill via `product-shots` if any of the 7 Quick Start fields is missing.
## Execution Procedure
```
run_workflow(brief) → variant[] + self_check
# Step 1: Platform + Objective
platform, format, ad_objective = brief.platform, brief.format, brief.ad_objective
load references/hard-constraints.md → dimensions / safe zones / text overlay / char limits
# Step 2: Industry + Brand
industry_bundle = load references/industry-style-rules.md[brief.industry]
if brief.brand_kit:
apply_brand_kit_overrides(industry_bundle, brief.brand_kit)
# brand colors → primary palette
# brand fonts → headlines
# brand logo → composed per platform safe-zone rules
# Step 3: Visual Asset Audit ("Available assets determine the creative ceiling")
asset_class = classify(brief.visual_assets)
∈ {has_product_photo, has_model_photo, has_logo_only, has_none}
# Step 4: Composition + Color (Industry × Objective cross-decision)
composition = select_composition_pattern(industry, ad_objective, asset_class)
# references/composition-patterns.md
color = pick_color_strategy(industry, brief.brand_kit)
# references/color-and-cta-strategy.md §Color
cta = pick_cta_strategy(ad_objective, platform)
# references/color-and-cta-strategy.md §CTA
# Step 5: Platform Adaptation Layer (inject platform-specific differences)
for plat in brief.platforms:
apply references/platform-style-profiles.md[plat]
apply references/multi-platform-output.md
sanitize_prompt(prompt, brief.copy)
# filter banned words; preserve user copy verbatim
# inject Required Positive Instruction (preserve native packaging text)
# Step 6: Self-Check Gate (Quality Check Before Delivery)
for variant in candidates:
pass = run_11_checks(variant, platform, industry, ad_objective, brief)
if not pass → regenerate
emit variant[] + self_check_score
```
## TOC
- [Step 1 — Platform + Objective](#step-1--platform--objective)
- [Step 2 — Industry + Brand](#step-2--industry--brand)
- [Step 3 — Visual Asset Audit](#step-3--visual-asset-audit)
- [Step 4 — Composition + Color](#step-4--composition--color)
- [Step 5 — Platform Adaptation Layer](#step-5--platform-adaptation-layer)
- [Step 6 — Self-Check Gate (11-Item Mandatory Checklist)](#step-6--self-check-gate-11-item-mandatory-checklist)
## Step 1 — Platform + Objective
Lock the (platform, format, ad_objective) tuple before any creative decision. Without it the safe zone is unknown, the char limit is unknown, the text overlay policy is unknown, and the visual hierarchy cannot be determined.
- Platform → load `references/hard-constraints.md §Platform Dimensions` for exact px and ratio.
- Format → resolves to one of the 21 platform×format combinations.
- Ad objective → drives information hierarchy (`references/ad-objective-rules.md`).
Refuse to proceed if the platform is ambiguous (e.g., "social media" is too vague — ask which platform).
## Step 2 — Industry + Brand
Match the user's industry to one of the presets in `references/industry-style-rules.md`. Load the bundle: color palette, composition pattern, typography, CTA tone, forbidden colors/elements.
If the user supplies a Brand Kit, integrate it as **hard constraints** that override industry defaults:
> If the user has Brand Kit (colors, fonts, logo), integrate it as hard constraints.
Brand Kit overrides:
- Brand colors → primary palette (industry colors become accent / supporting).
- Brand fonts → headlines (industry typography becomes secondary).
- Brand logo → composed per platform safe-zone rules (never in occluded regions).
The user's Brand Kit is parsed here when present — file path or inline fields. See `SKILL.md §Cross-Skill Notes (Brand Kit)`.
## Step 3 — Visual Asset Audit
> Available assets determine the creative ceiling.
Classify what the user has:
| Asset Class | Trigger | Composition Implication |
|---|---|---|
| `has_product_photo` | Product image only | Product Hero / Lifestyle In-Context / Flat Lay |
| `has_model_photo` | Person/model image | Face-Forward / Face-with-Product |
| `has_both` | Product + model | Face-with-Product (priority) |
| `has_logo_only` | Just brand logo | Bold Typography / Carousel Narrative |
| `has_none` | Nothing — AI generates everything | Bold Typography or industry-default lifestyle |
The asset audit is the input to Step 4's composition selection.
## Step 4 — Composition + Color
Use the **Industry × Objective × Asset** cross-decision logic in `references/composition-patterns.md` to pick the composition pattern. Then load the color strategy from `references/color-and-cta-strategy.md §Color Strategy` (industry color palette + max color count + temperature consistency + forbidden colors).
CTA strategy: pick CTA intensity from `references/color-and-cta-strategy.md §CTA Strategy` based on ad objective (low for Brand Awareness, high for Promotion / App Download).
The promotion/conversion objective has the strictest rules (discount number must dominate, must be in top 1/3, must use contrasting color, must include urgency cue) — see `references/ad-objective-rules.md`.
## Step 5 — Platform Adaptation Layer
For each target platform, inject platform-specific differences:
- **Tone & production level** — `references/platform-style-profiles.md` (TikTok = anti-ad / lo-fi; LinkedIn = professional authority; Pinterest = aspirational; etc.)
- **Text style** — TikTok native overlay; LinkedIn ≥60pt; Google Display zero text; YouTube big bold sans + drop shadow.
- **Refresh cadence** — TikTok 3-7 days; FB monthly; Pinterest weeks-months. See `references/multi-platform-output.md §Refresh Cadence`.
- **Prompt sanitation** — filter banned words from the generator prompt; inject the Required Positive Instruction. See `references/banned-words.md`.
Multi-platform output rules (consistent core message, platform-native style adaptation per plat) live in `references/multi-platform-output.md §Multi-Platform Output`.
## Step 6 — Self-Check Gate (11-Item Mandatory Checklist)
Every variant MUST pass all 11 checks before delivery. Any failure → revise prompt and regenerate.
```
QUALITY_CHECKLIST = [
"dimensions_correct", # dimensions exactly match the platform spec
"safe_zones_respected", # critical content avoids occluded zones
"text_rules_compliant", # complies with the platform's text policy
"one_dominant_focal_point", # one clear visual focal point
"cta_visible_appropriate", # CTA is visible and intensity matches the objective
"industry_visual_language", # matches the industry's visual language
"platform_tone_matched", # matches the platform's tone
"brand_kit_applied", # brand kit applied where supplied
"mobile_first_readable", # readable on a phone screen
"ad_objective_reflected", # ad objective is reflected
"user_content_preserved" # all user-specified content preserved verbatim
]
```
### Check 1 — Dimensions Correct
```
expected = PLATFORM_DIMENSIONS[platform]
actual = (creative.width, creative.height)
FAIL if actual != expected
```
### Check 2 — Safe Zones Respected
- IG Reel / FB Reel / YouTube Shorts → bottom 35% (y > height × 0.65) is forbidden zone for text / logo / CTA / product / face.
- TikTok → right 64px is forbidden zone for text / logo / CTA.
- Pinterest → logo must not sit in the bottom-right (search-icon overlap).
### Check 3 — Text Rules Compliant
- `google_display` → if any text overlay AND not natural product packaging text → FAIL ("CRITICAL: Google Display prohibits ALL text overlays").
- `instagram_feed` / `facebook_feed` / `twitter` → soft-warn if `text_area_ratio > 0.20` (won't fail, increases CPM).
- `tiktok` → FAIL if QR code or any other-platform watermark present.
- `linkedin` → FAIL if any text element has `font_size < 60pt` or `contrast_ratio < 4.5:1`.
### Check 4 — One Dominant Focal Point
```
focal_points = detect_focal_points(creative)
if len(focal_points) == 0: FAIL "No clear focal point detected"
if len(focal_points) > 1:
primary, secondary = focal_points[0], focal_points[1]
if primary.visual_weight / secondary.visual_weight < 1.5: FAIL "Multiple competing focal points"
```
### Check 5 — CTA Visible and Intensity-Appropriate
- Some platforms (`instagram_feed`, `facebook_feed`, `linkedin`) provide CTAs via the platform UI — image CTA not required.
- For `promotion_conversion` and `app_download` objectives, a visible image CTA is REQUIRED.
- CTA intensity must match the objective (low / mid / high — see `references/ad-objective-rules.md`).
### Check 6 — Industry Visual Language
- Dominant colors must NOT include any color in the industry's `forbidden_colors` list.
- Composition pattern should be in the industry's `recommended_patterns` list (warn if not).
### Check 7 — Platform Tone Matched
- TikTok: `production_quality > 0.7` → FAIL ("Production too polished, use lo-fi aesthetic").
- LinkedIn / Pinterest: `production_quality < 0.6` → FAIL.
- Production quality factors: lighting (studio vs natural), composition precision, color grading sophistication, camera stability.
### Check 8 — Brand Kit Applied
If user supplied a Brand Kit:
- Brand colors → must appear in the dominant color set.
- Brand fonts → must be used in at least one text element.
- Brand logo → must be present.
If no Brand Kit supplied → check N/A (auto-pass).
### Check 9 — Mobile-First Readable
```
mobile_width = 375 # iPhone SE reference
scale_factor = mobile_width / creative.width
for text in creative.text_elements:
scaled_pt = text.font_size * scale_factor
if scaled_pt < 12: FAIL "Text too small on mobile"
```
### Check 10 — Ad Objective Reflected
- `promotion_conversion` → must have a visible discount/offer; the discount number must be the largest text element on the canvas.
- `brand_awareness` → must have a prominent brand element (logo / brand name / signature visual).
- `lead_generation` → must include a trust signal (review / rating / testimonial / "as seen in").
- `app_download` → must include either a device mockup OR an app rating display.
### Check 11 — User Content Preserved (MUST-level)
This check is at the same priority as `dimensions_correct`. Every protected user phrase must be present:
- `brand_name` if supplied → must appear verbatim in the deliverable.
- `slogan` if supplied → must appear verbatim.
- `cta_message` if supplied → must appear verbatim.
- `price_or_offer` if supplied → must appear verbatim.
Any missing phrase = FAIL → regenerate. See `references/quick-start-and-fidelity.md §Section 5` for the fidelity rule.
### Self-Check Output
When all 11 checks pass, emit a self-check score block alongside the variant (per `references/failures-defaults-output.md §Output Format`). When any check fails, surface the failure reason to the agent and regenerate the prompt before re-running the gate.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |