Claude-skill-registry ad-review
Quality review ads before launch by verifying hook strength, checking proof elements, evaluating CTA effectiveness, and assessing visual quality and authenticity. Use as final check before launching creative or when reviewing team submissions.
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/ad-review" ~/.claude/skills/majiayu000-claude-skill-registry-ad-review && rm -rf "$T"
manifest:
skills/data/ad-review/SKILL.mdsource content
Ad Review
Quality check creative before launch.
Process
Step 1: Verify Hook Strength
Hook Checklist:
- Stops scroll in first 1-3 seconds
- Creates curiosity or emotional response
- Relevant to target avatar
- Doesn't reveal too much too soon
- Passes the "would I stop for this?" test
Hook Scoring:
| Criteria | Score (1-10) |
|---|---|
| Attention-grabbing | |
| Relevance to avatar | |
| Curiosity creation | |
| Clarity of message | |
| Average |
Red Flags:
- Generic opening ("Hey guys...")
- Product mentioned too early
- No pattern interrupt
- Looks like an ad immediately
Step 2: Check Proof Elements
Proof Checklist:
- Social proof present (testimonials, reviews)
- Authority elements (experts, credentials)
- Results/data shown
- Demonstration/proof of concept
- Believable, not too good to be true
Proof Quality:
| Element | Present | Quality | Authentic |
|---|---|---|---|
| Testimonial | Y/N | H/M/L | Y/N |
| Authority | Y/N | H/M/L | Y/N |
| Data/Results | Y/N | H/M/L | Y/N |
| Demo | Y/N | H/M/L | Y/N |
Red Flags:
- Claims without proof
- Fake-looking testimonials
- Over-the-top results
- Stock footage overuse
Step 3: Evaluate CTA Effectiveness
CTA Checklist:
- Clear call to action
- Easy to understand next step
- Creates urgency (without being fake)
- Risk reversal mentioned
- Placement is natural, not forced
CTA Types Present:
- Direct ("Click the link")
- Scarcity ("Sale ends...")
- Story-based (testimonial leading to action)
Red Flags:
- No clear CTA
- CTA too early (kills the story)
- Fake scarcity
- Confusing next steps
Step 4: Assess Visual Quality & Authenticity
Visual Checklist:
- Quality appropriate for platform (native feel)
- Not over-produced
- Audio clear and understandable
- Text overlays readable
- Mobile-optimized
Authenticity Check:
- Feels genuine, not scripted
- Reactions appear real
- Setting is believable
- Person is relatable to avatar
- Would pass the "comments won't call it fake" test
Red Flags:
- Too polished/professional
- Scripted delivery obvious
- Mismatched audio/video quality
- Actor clearly not real user
- Generic stock footage feel
Step 5: Output Approval Decision
## AD REVIEW: [Creative Name/ID] ### QUICK DECISION **Status:** [APPROVED / REVISE / REJECT] **Overall Score:** [X/100] **Priority Issues:** [#] --- ### SCORES BY CATEGORY | Category | Score | Status | |----------|-------|--------| | Hook Strength | X/25 | [Pass/Needs Work] | | Proof Elements | X/25 | [Pass/Needs Work] | | CTA Effectiveness | X/25 | [Pass/Needs Work] | | Visual/Authenticity | X/25 | [Pass/Needs Work] | | **TOTAL** | X/100 | | --- ### HOOK REVIEW **Current Hook:** "[Description/transcript]" **Strengths:** - [What's working] **Issues:** - [Problem 1] - [Problem 2] **Recommendation:** [Keep as is / Modify / Replace] **If Modify:** [Specific suggestion] --- ### PROOF REVIEW **Proof Elements Present:** - [Element 1]: [Assessment] - [Element 2]: [Assessment] **Missing:** - [What should be added] **Authenticity Assessment:** [Does proof feel real and believable?] **Recommendation:** [Strengthen / Add / Acceptable] --- ### CTA REVIEW **Current CTA:** "[Description/transcript]" **Placement:** [Too early / Just right / Too late] **Clarity:** [Clear / Confusing] **Urgency:** [Natural / Forced / Missing] **Recommendation:** [Keep / Modify / Add urgency] --- ### VISUAL/AUTHENTICITY REVIEW **Production Quality:** [UGC-native / Slightly polished / Over-produced] **Audio Quality:** [Good / Acceptable / Needs work] **Authenticity:** [Genuine / Borderline / Fake-feeling] **Issues:** - [Problem 1] - [Problem 2] **Recommendation:** [Approve / Re-edit / Re-shoot] --- ### PLATFORM FIT **Best For:** - [ ] Facebook Feed - [ ] Instagram Reels - [ ] TikTok - [ ] YouTube - [ ] Native **Platform-Specific Notes:** - [Any platform adjustments needed] --- ### COMPLIANCE CHECK - [ ] Claims supportable - [ ] Disclosures present (if needed) - [ ] No policy violations --- ### ACTION ITEMS **Required Before Launch:** 1. [ ] [Must fix] 2. [ ] [Must fix] **Recommended:** 1. [ ] [Should improve] **Optional:** 1. [ ] [Nice to have] --- ### FINAL DECISION **Approved for launch:** [YES / NO / WITH CHANGES] **Reviewer:** [Name] **Date:** [Date] **Notes:** [Additional comments] --- ### COMPARISON TO WINNERS How does this compare to top performers? - Hook style: [Similar / Different] - Proof approach: [Similar / Different] - Energy/tone: [Similar / Different] - Prediction: [High potential / Medium / Low]
Review Standards (LeadsIcon)
Authenticity Test: "Would comments call this fake?"
- If yes → needs work
- If unsure → probably needs work
- If no → might work
Platform Native Test: "Does this look like content or an ad?"
- Content = good
- Ad = bad
The Real Test: Would the person in the video actually send this to a friend recommending the product?
Source: LeadsIcon, Kamal FounderAds