ads-creative
Cross-platform creative quality audit covering ad copy, video, image, and format diversity across all platforms. Detects creative fatigue, evaluates platform-native compliance, and provides production priorities. Collects brand context and creates branding.md if missing. Uses infographic-v2 for generating ad creatives. Use when user says "creative audit", "ad creative", "creative fatigue", "ad copy", "ad design", or "creative review".
适合你,如果你需要跨平台检查广告素材质量并提升投放效果
用别的 agent?下载 .zip 解压,把文件夹放进它的技能目录
~/.claude/skills/(项目级 .claude/skills/)~/.codex/skills/npx oh-my-skill add naveedharri/benai-skills/ads-creativecurl -fsSL https://oh-my-skill.com/install.sh | bash -s -- naveedharri/benai-skills/ads-creativenpx oh-my-skill verify naveedharri/benai-skills/ads-creative怎么用
技能原文 SKILL.md
Cross-Platform Creative Quality Audit
Process
- Brand context check — look for
./branding.md. If not found, run the Brand Info Collection flow below before proceeding. - Collect creative assets or performance data from active platforms
- Read
ads/references/platform-specs.mdfor creative specifications - Read
ads/references/creative-volume.mdfor per-platform volume and refresh requirements - Read
ads/references/benchmarks.mdfor CTR/engagement benchmarks - Read
ads/references/scoring-system.mdfor weighted scoring algorithm - Evaluate creative quality per platform
- Run Brand Consistency Audit using branding.md (advisory, not scored)
- Assess cross-platform creative consistency
- Generate production priority recommendations
- Offer to generate new creatives using
/infographic-v2for any flagged gaps
Brand Info Collection
Run this at the start if ./branding.md does not exist in the project root.
Goal: Extract the brand's actual styling, copy, and identity — primarily from their website — to personalize the creative audit and enable on-brand creative generation via infographic-v2.
Step 1: Ask for brand name and website
"Before I audit your creatives, I need your brand context to check consistency. What's your brand name and website URL?"
Ask for the brand name and website URL first. If the user provides a URL, immediately run the website extraction (step 2) before asking any further questions.
Step 2: Website extraction (when URL provided)
Do not guess or assume anything. Extract the actual values from the live website.
2a — Fetch the homepage
Use WebFetch on the homepage URL. Extract ALL of the following:
Visual identity:
- All colors used prominently: backgrounds, buttons, headers, links, text, borders, footer. Report exact hex values. Identify which is primary (most prominent brand color, usually on CTAs and headers), secondary, accent, background, and text color.
- Font families: read the actual
font-familydeclarations used on headings and body text. Look in CSS,<link>tags for Google Fonts / Typekit / custom font URLs, and inline styles. Report the exact font names (e.g., "Plus Jakarta Sans", "DM Sans", "Geist"), not generic fallbacks like "sans-serif". - Logo: describe the logo from the header/nav. Note if it's text-based, icon-based, or both.
- Visual style: describe what you observe — minimal with whitespace? Bold with saturated colors? Dark mode? Gradient-heavy? Illustration-driven? Corporate?
- Button styles: rounded, square, pill-shaped? What color? What text color on buttons?
- Spacing and density: tight and information-dense, or airy and spacious?
Copy and messaging:
- Tagline / slogan from the hero section
- Hero headline: the main H1 or hero text
- Value proposition: what they do and why it matters
- Key benefits: any benefit sections, feature highlights, selling points
- Social proof: testimonials, customer logos, metrics, awards, trust badges, star ratings
- CTAs: every call-to-action button text on the page
- Navigation items: main nav labels reveal product/service categories
Brand voice signals:
- Tone of the copy: formal or casual? Technical or plain? Playful or serious?
- Person: first person ("We help...") or second person ("You can...")?
- Specific words/phrases: any repeated brand language, trademarked terms
2b — Fetch additional pages (if needed)
If the homepage left gaps, also fetch:
/aboutor/about-us— company story, mission, team, more voice signals/pricing— pricing model (free trial, freemium, tiers, contact-us)
Only fetch if there are genuine gaps. Don't fetch for the sake of it.
2c — Present extraction to user
Show everything extracted clearly:
Here's what I found on [website]:
Brand: [name]
Tagline: "[extracted tagline]"
Industry: [detected industry]
Business model: [detected model]
Colors:
Primary: [name] #[hex] — used on [CTAs, headers, etc.]
Secondary: [name] #[hex] — used on [subheadings, etc.]
Accent: [name] #[hex] — used on [highlights, badges, etc.]
Background: #[hex]
Text: #[hex]
Fonts:
Headings: [exact font name]
Body: [exact font name]
Visual style: [description of what you observed]
Button style: [pill/rounded/square], [color], [hover behavior if visible]
Voice: [formal/casual], [serious/playful], [technical/simple]
Hero: "[exact hero headline]"
Value prop: "[extracted value proposition]"
CTAs found: "[CTA 1]", "[CTA 2]", "[CTA 3]"
Social proof: [what's shown — logos, metrics, testimonials]
Does this look right? Anything to correct or add?
The user confirms or corrects. Only after confirmation, proceed with remaining questions.
Step 3: Fill gaps with questions
After website extraction, only ask about things that could not be determined from the website. Skip any section where the website gave a clear answer.
Likely still needed (websites rarely cover these):
- Target audience: who they sell to (job title/role, age), pain points, buying triggers, where they spend time online
- Ad-specific guidelines: confirm extracted CTAs, pricing in ads?, required elements (trust badges, ratings), anything forbidden
- Brand voice refinement (only if ambiguous from the site): words they always/never use
Do NOT re-ask about colors, fonts, tagline, visual style, or value proposition if the website extraction already captured them.
Step 4: No-URL fallback
If the user does not provide a website URL, collect everything manually:
- Brand colors — primary, secondary, accent (hex codes)
- Typography — headline and body fonts
- Visual style — clean/minimal, bold/colorful, professional, playful, premium
- Tagline / value proposition
- Brand voice — formal vs casual (1-10), serious vs playful, technical vs simple
- Target audience — who they sell to, main pain points
- Approved CTAs
- Anything forbidden in ads
Step 5: Save branding.md
Write ./branding.md in the project root using the canonical format from ads/references/brand-context.md.
Every value in branding.md must come from the website extraction or the user's direct answers. Never fill in defaults or placeholders. If a field has no data, leave it blank or omit the section.
Validate before writing:
- Brand name is not empty
- At least one color with hex value exists
- At least one audience description exists
- Business model is identified
Step 6: Confirm and continue
Brand context saved: ./branding.md Brand: [name] Colors: [primary hex] [secondary hex] [accent hex] Fonts: [heading font] / [body font] Voice: [tone summary] Audience: [persona summary] Proceeding with creative audit...
Then continue with the normal creative audit flow (step 2 onward in the Process section).
Per-Platform Assessment
Google Ads Creative
- RSA: ≥8 unique headlines, ≥3 descriptions per ad group
- RSA ad strength: "Good" or "Excellent"
- Pin usage: minimal and strategic (over-pinning kills RSA flexibility)
- Extensions: sitelinks (≥4), callouts (≥4), structured snippets, image
- PMax asset groups: text + image + video + optional product feed
- YouTube: video quality, hook, subtitles (see ads-youtube sub-skill)
Meta Ads Creative
- Format diversity: ≥3 formats active (image, video, carousel, collection)
- Creative volume: Standard ≥5 per ad set (ideal 5-8 for Andromeda); ASC ≥50 (ideal 50-150)
- Fatigue detection: CTR declining >20% over 14 days = FAIL; top brands refresh every 10.4 days
- Video length: 15s max Stories/Reels, 30s max Feed
- UGC/testimonial content tested
- Advantage+ Creative enhancements enabled
- Headline under 40 chars, primary text under 125 chars
LinkedIn Creative
- Thought Leader Ads active, ≥30% budget for B2B
- Format diversity: ≥2 formats tested (single image, carousel, video, document)
- Video ads tested
- Creative refresh: every 4-6 weeks
- Professional tone appropriate for platform
TikTok Creative
- ≥6 creatives per ad group (Critical; top performers: 10-20/campaign; Smart+ launch: ≥4)
- All video 9:16 vertical 1080x1920 (non-negotiable)
- Native-looking content (not corporate)
- Hook in first 1-2 seconds
- Fatigue onset: 3-5 days; typical lifespan: 7-10 days; refresh every 10-14 days
- No creative active >7 days with declining CTR; hard ceiling: 30 days regardless
- Spark Ads tested (~3% CTR vs ~2% standard)
- Sound-on optimization (never silent)
- Safe zone compliance: X:40-940, Y:150-1470
- Trending audio used
Microsoft Creative
- RSA: ≥8 headlines, ≥3 descriptions
- Multimedia Ads tested (unique rich format)
- Ad copy optimized for Bing demographics (older, higher income, professional)
- Action Extension utilized (unique to Microsoft)
- Filter Link Extension tested
Creative Fatigue Detection
Signals of Fatigue
| Signal | Threshold | Action | |--------|-----------|--------| | CTR declining | >20% over 14 days | Refresh creative | | Frequency (Meta) | >5.0 prospecting, >12.0 retargeting | New audience or creative | | Watch time declining (TikTok) | <3s average | New hook needed | | QS declining (Google) | Drop of 2+ points | Refresh ad copy | | Engagement rate drop | >30% decline | Full creative overhaul |
Refresh Cadence by Platform
| Platform | Recommended Refresh | |----------|-------------------| | Google Search | Every 8-12 weeks | | Meta | Every 10-14 days (top brands avg 10.4 days); hard ceiling 90 days | | LinkedIn | Every 2-4 weeks; 80/20 winner/testing split | | TikTok | Every 10-14 days (fatigue onset 3-5 days, lifespan 7-10 days); hard ceiling 30 days | | Microsoft | Every 8-12 weeks | | YouTube | Every 4-8 weeks |
Format Diversity Matrix
Evaluate which formats are active per platform:
| Format | Google | Meta | LinkedIn | TikTok | Microsoft | |--------|--------|------|----------|--------|-----------| | Static Image | RSA image ext | ✅ | ✅ | ❌ | Multimedia | | Video | YouTube, PMax | ✅ | ✅ | ✅ (required) | ❌ | | Carousel | ❌ | ✅ | ✅ | ❌ | ❌ | | Collection | ❌ | ✅ | ❌ | ❌ | ❌ | | Document | ❌ | ❌ | ✅ | ❌ | ❌ | | Shopping | PMax, Shopping | Catalog | ❌ | Shop | Shopping |
Universal Creative Best Practices
Cross-Platform Safe Zone
- 900x1000px usable area works across all vertical placements
- Keep critical elements centered and within safe margins
- Test on mobile devices (75%+ of ad impressions are mobile)
Ad Copy Principles
- Lead with benefit, not feature
- Include clear CTA (what should they do next?)
- Match ad message to landing page (message match)
- Use numbers and specifics over vague claims
- Test emotional vs rational appeals
Video Production Standards
- H.264 codec, AAC audio, MP4 container
- Minimum 720p (1080p preferred)
- Subtitles/captions always (accessibility + sound-off viewing)
- Brand mention within first 5s (awareness) or at CTA (performance)
Brand Consistency Audit (Advisory)
Using ./branding.md, evaluate these dimensions. Results are advisory only — they appear as "Brand Consistency Notes" in the report but are NOT included in the health score.
| Dimension | What to Check | |-----------|--------------| | Color compliance | Ad visuals use brand primary/secondary colors (hex match) | | Typography | Headlines and body text match brand font guidelines | | Tone alignment | Ad copy matches brand voice dimensions (formal/casual, etc.) | | CTA consistency | CTAs match approved CTA list from branding.md | | Visual style | Photography/illustration style matches brand guidelines | | Forbidden elements | No forbidden words or elements present in ads | | Audience match | Messaging addresses documented persona pain points |
Creative Generation with Infographic v2
After the audit, when the production priority list identifies creatives to produce (static images, infographics, data visuals for ads), offer to generate them using /infographic-v2.
Flow:
- Identify the highest-priority static creative gaps from the audit
- Ask: "I can generate ad creatives for the top gaps using the infographic tool. Want me to start with [highest priority gap]?"
- If yes, invoke
/infographic-v2— it will pick up./branding.mdfor brand-consistent output - The infographic skill handles concept selection, visualization approach, and generation
- After generation, return to the next creative gap or wrap up
What infographic-v2 uses from branding.md:
- Brand colors → color palette for the infographic
- Typography → font selection
- Visual style → style anchors
- Brand voice → headline and copy tone
- Approved CTAs → CTA text in ad creatives
Output
Creative Quality Report
Cross-Platform Creative Health Google: ████████░░ X/X checks passing Meta: ██████████ X/X checks passing LinkedIn: ███████░░░ X/X checks passing TikTok: █████░░░░░ X/X checks passing Microsoft: ████████░░ X/X checks passing
Deliverables
CREATIVE-AUDIT-REPORT.md— Per-platform creative assessment- Fatigue alerts (any creative past refresh cadence)
- Format diversity gaps per platform
- Production priority list (most impactful creative to produce next)
- Quick Wins (format conversions, CTA changes, Spark Ads setup)
- Generated ad creatives via infographic-v2 (if user opts in)