What makes a good copywriting skill
Copywriting is one of the areas where a skill's value is easiest to verify: does the output sound like your brand, and does it hit the format constraints of the channel it's for? The weak version of this category is a skill that just says "writes persuasive copy" and leaves voice and structure entirely to the model's default instincts. The strong version either enforces guidelines you already have, or is scoped tightly enough to a channel (cold email, paid ads, release notes) that its structural conventions do real work.
A few things worth checking before you install one:
- Does it need your brand guidelines as an input, or does it invent a voice? Brand Voice Enforcement and Brand Review both work by applying guidelines you supply. That's the more reliable pattern, a skill that improvises a "brand voice" from a one-line prompt is really just guessing.
- Is it scoped to a channel with real constraints? Cold Email Writing, Ad Creative and LinkedIn Post Formatter each target a channel with actual format rules, reply-earning subject lines, platform-specific ad review requirements, LinkedIn's lack of rich text. That specificity is what makes the output usable without heavy editing.
- Does it separate drafting from review? Draft Content generates; Brand Review audits what already exists. Pairing a drafting skill with a review skill catches more issues than expecting one skill to do both well.
- Does it require script execution? Copywriting and Ad Creative both run Python; the rest of this list is pure-instruction and platform-agnostic.
The order below starts with the most general-purpose copywriting skills (the ones useful across almost any page or channel) then narrows toward skills scoped to a single format or a single emerging discipline.
The 10 best copywriting skills for AI agents right now
Copywriting
A general-purpose marketing copy skill for any page type, homepage, landing pages, pricing pages, feature pages, about pages, product pages. It works by first gathering context about the product and audience, then applying principles of clarity, specificity and customer language rather than generic persuasive-writing tropes.
Who it's for: anyone writing or rewriting web page copy who wants a structured process rather than a single-shot draft. Caveat: the author notes it's not built for technical documentation or creative writing tasks requiring a different register. Stay within marketing copy for web pages, and reach for a documentation-specific skill elsewhere. Requires Python and shell access. From alirezarezvani/claude-skills (24,000 stars, MIT).
Draft Content
Drafts blog posts, social media, email newsletters, landing pages, press releases and case studies with channel-specific formatting and SEO recommendations, including headline and subject-line options and adaptation for a specific platform, audience and brand voice.
Who it's for: marketers and content teams who need a first draft across several channels without switching tools for each one. Caveat: the author flags it as a weaker fit for highly specialized or technical content requiring deep subject-matter expertise beyond general marketing principles. It drafts competently across formats, but it isn't a domain expert. No script execution required. From anthropics/knowledge-work-plugins (23,000 stars, Apache-2.0).
Brand Voice Enforcement
Applies existing brand guidelines to content creation (emails, proposals, pitch decks, LinkedIn posts, presentations, Slack messages, sales content) by loading your guidelines first and then checking that the requested content's style and messaging match before producing output.
Who it's for: teams with documented brand guidelines who want them applied consistently across every piece of content an agent drafts, not just the ones someone remembers to check manually. Caveat: the author is explicit it isn't for generating brand guidelines from scratch or discovering existing brand materials. You need the guidelines to already exist somewhere the skill can read them. No script execution required. From anthropics/knowledge-work-plugins (23,000 stars, Apache-2.0).
Brand Review
Reviews content against your brand voice, style guide and messaging pillars, flagging deviations by severity with specific before/after fixes, plus screening for unsubstantiated claims, missing disclaimers, and other legal flags before something ships.
Who it's for: teams that want a final check before content publishes, especially content produced by multiple authors or channels where consistency tends to drift. Caveat: the author notes it may not suit informal content that doesn't need strict brand adherence, or non-marketing content types entirely. It's built for the final-mile check on content that matters. No script execution required. From anthropics/knowledge-work-plugins (23,000 stars, Apache-2.0).
UX Copy
Writes or reviews microcopy (button labels, error messages, empty states, calls to action, confirmation dialogs, onboarding text) triggered by prompts like "write copy for" or "what should this button say?", producing tailored suggestions once given context, user state, tone and constraints.
Who it's for: product and design teams who need short, precise interface copy rather than long-form marketing content. Caveat: the author is clear it isn't built for long-form content or detailed technical documentation. This is a microcopy specialist, and asking it to draft a blog post is outside its scope. No script execution required. From anthropics/knowledge-work-plugins (23,000 stars, Apache-2.0).
Ad Creative
Generates, iterates and scales ad copy for paid advertising across Google, Meta, LinkedIn and Twitter (headlines, ad variations, bulk creative, copy validation against platform review requirements) operating in three modes: generating from scratch, iterating on existing performance data, and scaling variations of what's already working.
Who it's for: performance marketers who need creative production specifically, separate from campaign strategy or budget decisions. Caveat: the author is explicit this is pure creative production, not campaign strategy or budget allocation, pair it with a separate strategy process rather than expecting it to plan a campaign. Requires Python. From alirezarezvani/claude-skills (24,000 stars, MIT).
Cold Email Writing
Writes B2B cold emails and follow-up sequences aimed at earning replies from prospects with no prior relationship, personalized opening lines, subject lines, calls to action, and multi-touch follow-up sequences, guided by questions about target audience, desired outcome, and the sender's unique value.
Who it's for: SDRs and marketers doing outbound prospecting who need email copy structured around getting a reply, not just sounding professional. Caveat: the author notes it's a weaker fit for extensive product pitches or contacting people already familiar with your offering. It's built specifically for cold, no-relationship outreach. No script execution required. From sickn33/agentic-awesome-skills (45,000 stars, MIT).
Release Announcement
Writes release communications (changelog entries, blog posts, in-app notes, social posts) that lead with user impact, name the affected audience, and include upgrade or migration steps without filler, across whichever channel the release needs to reach.
Who it's for: product teams and developers who need to communicate a new version, feature or fix clearly across more than one channel from a single source of truth. Caveat: the author flags it as a poor fit for internal-only changes or releases still in active development that aren't ready for public announcement. It's built for the public-facing moment, not internal status updates. No script execution required. From paperclipai/paperclip (76,000 stars, MIT).
LinkedIn Post Formatter
Formats and drafts LinkedIn posts using Unicode Mathematical Alphanumeric Symbols to simulate bold, italic and bold-italic text, working around LinkedIn's lack of native rich-text formatting, plus visual separators, structured sections and engagement-optimized patterns.
Who it's for: anyone drafting thought-leadership or company posts who wants visual structure on a platform that otherwise renders everything as plain text. Caveat: the author notes it's not a full content-creation tool and has no graphic design capability. It structures and formats copy you've already drafted rather than originating the underlying argument. No script execution required. From github/awesome-copilot (38,000 stars, MIT).
Answer Engine Optimization
Optimizes content to be cited by AI language models (ChatGPT, Perplexity, Claude, Gemini, Mistral) as an authoritative source, distinct from traditional SEO because it targets citation in LLM-generated answers rather than search-result rankings, auditing content for E-E-A-T signals and tracking which pages get cited by which models.
Who it's for: content and copy teams thinking ahead to an AI-first search landscape who want their pages to be the ones an LLM actually quotes. Caveat: the author notes it's a poor fit for traditional SEO tasks focused purely on click-through rate, or for content without factual claims suitable for citation. This is a genuinely different discipline from SEO, not a replacement for it. Requires Python and shell access. From alirezarezvani/claude-skills (24,000 stars, MIT).
How to install any of these skills
Every skill above is a plain folder containing a SKILL.md file. Nothing to license, nothing to register. The fastest install path is the cross-agent skills CLI, run once from your project root:
npx skills add alirezarezvani/claude-skills/copywriting --agent claude-code
Swap the path for any installCoordinate listed above, and swap --agent claude-code for the agent you're targeting, the same command pattern installs into Codex CLI, Cursor, Antigravity and others, with only the destination folder changing.
To install by hand instead, clone or copy the folder into your agent's skills directory (~/.claude/skills/ for personal-scope Claude Code, .claude/skills/ for project scope) then start a new session. Full walkthroughs are in our guides on installing skills in Claude Code and installing skills in Codex CLI. Marketing teams that don't work from a terminal can install the same skills through Claude Cowork instead, and share them with the wider team on Team and Enterprise plans.
Combining several of these into one workflow
Copywriting skills tend to work best chained, since drafting, brand-checking and channel formatting are naturally separate steps:
- Draft Content → Brand Voice Enforcement → Brand Review. Generate the first draft, apply brand guidelines during that draft rather than after, then run a final audit before it ships. This is the closest thing to a full pipeline in this list, and all three are from the same author.
- Cold Email Writing + Ad Creative for a coordinated outbound push. Both are channel-specific production skills rather than strategy tools, so running them side by side for a single campaign keeps messaging consistent across email and paid channels without duplicating effort.
- UX Copy alongside Copywriting. Use Copywriting for the page-level narrative and UX Copy for the buttons, error states and confirmation dialogs on that same page. They operate at different zoom levels and rarely conflict.
Running two general-purpose drafting skills against the same brief tends to produce redundant or conflicting output. Pick the one whose scope actually matches the channel you're writing for, rather than layering broad skills on top of each other.
Where these skills fit in the broader ecosystem
This is a curated selection based on how clearly each skill's description names a specific writing task, how actively its source repository is maintained, and how well it separates drafting from review, not a benchmark or test result. Browse the current, full list at the copywriting category page, the broader writing category, or check all skills for something more specific.
If you're new to the format, what agent skills are and the SKILL.md format explained cover the basics. And if your brand voice has never been written into a skill, writing your own is often the fastest path to copy that actually sounds like you, most of the guideline-driven skills above started as exactly that adaptation.
