copy-pass is a senior-writer review skill for improving marketing and product copy before it reaches a final human editor. It gives an AI agent a structured way to check whether a draft is clear, persuasive, specific, and ready for a senior writer to review.
Use it after the first draft and before final humanisation. The skill does not try to make copy sound more natural. Its job is to strengthen the message: the promise, hook, structure, objection handling, CTA, emotional trigger, and platform fit.
Best paired with humaniser, which should run after copy-pass to remove AI tells and make the final copy sound more human. For the conversational check, use care-review.
copy-pass condenses reusable frameworks from 19 marketing, advertising, copywriting, argumentation, and behaviour-change books/playbooks. See references/index.md for the full source list and references/source-frameworks/ for the extracted framework notes.
- Weak hooks that do not earn attention
- CTAs that are vague, repeated, or trying to do too much
- Copy that explains features but does not make the value feel urgent
- Missing emotional triggers or buyer motivation
- Drafts that are clear but not persuasive
- Platform mismatch, especially when the same copy is reused across email, social, landing pages, SMS, and ads
- Objections that a reader will have but the draft ignores
This tool is part of a three-skill writing stack. Each tool does one job well.
| Tool | What it does | When to use it |
|---|---|---|
| copy-pass | Strengthens persuasion: hooks, CTAs, objections, emotional triggers, platform fit | Before a senior writer reviews copy. Not for final cleanup. |
| care-review | Checks whether a message is conversational, actionable, richer than asked, and ends with a reason to reply | Before any customer-facing message goes out. |
| humaniser | Removes AI writing patterns and makes text sound natural | After copy pass. Final voice pass before publishing. |
- Your copy sounds like AI wrote it → run humaniser after
- Your text needs a conversational check → run care-review after
- You only need persuasion work → stay here, skip care-review and humaniser
Draft → copy-pass → care-review → humaniser → publish / send
Run copy-pass before humaniser. copy-pass adds persuasion structure, and that can reintroduce patterns that sound like AI writing. humaniser should clean the final draft after the persuasive edits are in place. care-review runs between them because persuasion edits can also make text sound less conversational.
git clone https://github.com/kastrah/copy-pass.git ~/.claude/skills/copy-passCopy SKILL.md into your agent's skills directory, or paste the contents into your system instructions. Keep the references/ folder beside it so the agent can load the right checklist, CTA guide, hook guide, and objection-handling notes when needed.
Ask your agent to run copy-pass on a draft:
Run copy-pass on this landing page copy before we send it to the senior writer.
Or:
Review this campaign copy using copy-pass. Strengthen the hook, CTA, objections, and emotional trigger, then give me the revised version.
For each review, the agent should give you:
- A short diagnosis of what is weak and why
- The specific reference areas used, such as hooks, CTAs, objections, platform fit, or persuasion formulas
- A revised draft
- Notes for the senior writer, only where judgment is still needed
The final copy should not include internal audit notes. Keep production notes in the review summary, not inside the public-facing copy.
references/review-checklist.md— 18-point review checklistreferences/principles.md— core copy principlesreferences/headline-hooks-openers.md— hooks and openersreferences/cta-framework.md— CTA hierarchy and action clarityreferences/triggers-and-motives.md— emotional triggers and buyer motivationsreferences/persuasion-formulas.md— structure formulas for different copy typesreferences/objection-handling.md— finding and resolving reader objectionsreferences/social-platform-guide.md— platform-specific execution notesreferences/index.md— quick map of the reference system
MIT