ai-marketing-claude-code-skills: marketing frameworks packaged as Claude Code skills
Marketing frameworks that AI actually executes. Designed for Claude Code, but could be used for OpenClaw, etc.
At a glance
- What is it?
- A repository of 23 free agent skills that drop marketing frameworks into Claude Code's skills folder, plus 10 paid ones. Useful if you already work in Claude Code, thinner if you expect the skills to fetch their own data.
- Who is it for?
- Adopt it if you already run Claude Code and want positioning, content and audit frameworks loaded automatically instead of retyping prompts. Skip it if you need research skills that pull live data without a separate search tool, or if you cannot work in the Claude Code skills folder at all.
- Can I use it commercially?
- Not without permission. GitHub finds no licence file in the repository, and without a licence all rights are reserved by default: you may read the code but not reuse it. Check the README, or ask the authors, before using it.
- Is it still maintained?
- Activity is slowing. The repository last received commits 6 months ago.
- What is it written in?
- Mainly Shell, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on October 1, 2026, and from our analysis. They are not legal advice.
Editorial analysis
The problem: marketing advice that never gets applied twice the same way
Most marketing material is written for a human to read once and then improvise from. The repository's own framing puts it bluntly: traditional content is something you read, apply, and then forget half of. The pitch here is that the framework lives in a file the agent loads, so the same questions and output format come back every time.
That is a real distinction, not a marketing one. A prompt you type from memory drifts. A skill file is versioned, diffable, and identical across sessions and machines. If you have ever asked an assistant for a homepage audit and gotten a different shape of answer on Tuesday than on Monday, the appeal is obvious.
The audience is narrow and specific: founders, marketers and consultants who already use Claude Code as their working environment. The README states the skills are designed for Claude Code and could be used for OpenClaw and similar agents. If you do not run an agent that reads skill files, nothing here executes.
What a skill actually is in this repo
Each skill is a folder with a SKILL.md file at its root. The repository layout confirms this: positioning-basics/, homepage-audit/, last30days/, voice-extractor/ and the rest each sit at the top level, and the README links directly to the SKILL.md inside each one.
The README describes skills as an open standard for packaging expertise as instructions an agent can follow, and points to agentskills.ai for the specification. Practically, a skill file contains the framework, the questions to ask, and the output format. The agent reads it when your prompt touches the skill's topic, which the README says happens automatically once the files are in place.
Version 3.1 added execution modes. The README states that all 19 free skills now support quick, standard and deep modes, so the same skill can return a fast answer or a longer audit without extra prompting. The repository also carries a SKILL-MODE-PATTERN.md file at the top level, which suggests the mode convention is documented separately from any single skill.
The skills are grouped by job: strategy and positioning, content and authority, research and intelligence, conversion and sales. Each README entry lists trigger phrases. Positioning Basics responds to "Who is this for?" and "What makes us different?". Homepage Audit responds to "Why isn't my page converting?". That trigger list is the interface. There is no CLI, no server, and no API.
Installing the skills and running a first audit
The README gives a two-step install: clone the repository, then copy its contents into Claude Code's skills folder. The commands below are copied from the README exactly.
git clone https://github.com/BrianRWagner/ai-marketing-claude-code-skills.git
mkdir -p ~/.claude/skills
cp -r ai-marketing-claude-code-skills/* ~/.claude/skills/Note what the copy does. It takes everything at the repository root, which includes not only the skill folders but also README.md, CHANGELOG.md, SKILL-MODE-PATTERN.md and the scripts/ directory. If you would rather keep the skills folder clean, copy individual skill directories instead of the wildcard. The README does not discuss that option.
After the copy, the README says Claude Code will use the skills automatically when you mention related topics. To exercise one, open Claude Code in any project and ask a question that matches a trigger phrase. For a homepage review:
claudeThen, inside the session, prompt with something close to the README's trigger wording, for example: review my homepage and tell me why it isn't converting. The expected behaviour is that the agent loads homepage-audit/SKILL.md and returns feedback in the format that file defines. The README describes the result as actionable feedback in minutes.
One thing the install section does not cover: there is no uninstall or update procedure. Re-running the copy overwrites files, but the README does not state whether that is the intended upgrade path. Verify against CHANGELOG.md before you overwrite a skill you have edited locally.
Where the free set stops and the paid set starts
The README's badge says 23 free skills plus 10 pro, while the v3.1 note says all 19 free skills support the new modes. Those two numbers do not agree, and the README does not reconcile them. Treat the count as approximate until you list the directories yourself.
More consequential is what sits behind a paywall. The Voice Extractor entry links to a nine dollar premium version on Claw Mart. The free version is in the repository; the paid one is not described in terms of what it adds. If voice extraction is the reason you came, that gap matters.
Several research skills are also thinner than their names suggest. Reddit Insights promises semantic AI search over Reddit content, but the README does not name the search backend, an API key, or a rate limit. Last 30 Days Research claims coverage of Reddit, X and the web in seven minutes versus two hours of manual work. No methodology is given for that comparison. Social Card Generator is the exception: the README explicitly states it has no API dependency.
Read this as a packaging project, not a data project. The value is in the frameworks and their output formats. Where a skill needs live data, the README leaves the retrieval mechanism to the agent or to you.
The honest limits: model-dependent, Claude Code-shaped, and thin on rollback
The first limitation is structural. These skills are instructions, not software. Their output depends entirely on the model reading them. The README's claim that the agent applies the framework "every time. Perfectly" is a promise about consistency of process, not about quality of result, and nothing in the repository enforces it.
The second is the install path. Copying into ~/.claude/skills means the skills are global to your machine. There is no per-project scoping described, so a skill loaded for client work is also loaded for unrelated repositories. The README does not document how to disable a skill selectively or how to roll back after an update.
The third is the compatibility claim. The README says the skills also work with OpenClaw, Copilot and Cursor, but the installation instructions cover Claude Code only. If you are on one of the other agents, you are translating the setup yourself.
Finally, the repository has no releases. The README points to CHANGELOG.md and a March 2026 v3.1 note, but there are no tagged versions to pin. If you need reproducible skill content across a team, you are copying a moving target.
How this differs from a prompt library or a marketing SaaS tool
The obvious alternative is a prompt library: a Notion page, a Gist, a shared doc of prompts you paste in. The difference is where the text lives. A prompt library requires you to select the right prompt and paste it. A skill sits in the agent's skills folder and is selected by the agent when your phrasing matches. The trade-off is control: with a prompt library you always know which prompt ran, and with a skill you are trusting the trigger match.
The second alternative is a marketing platform with its own AI features, such as a social scheduling tool or an SEO suite with a built-in assistant. Those come with data: your post history, your rankings, your traffic. This repository ships frameworks and no data. It will not tell you what your competitors rank for, and the README does not claim it will. If your problem is measurement, you want the platform. If your problem is that you keep producing inconsistent positioning and content, the framework is the part you were missing.
A third comparison is worth making inside the repository itself. Skills like Homepage Audit and Positioning Basics need nothing but your input and the model. Skills like Last 30 Days Research and Reddit Insights need retrieval. The first group is self-contained; the second depends on tooling the README does not specify. Judge them separately.
Licence, maintenance and the cost of keeping skills current
The README carries an MIT licence badge linking to opensource.org, and the repository's own licence field is unset. Those two facts disagree. If you need certainty about terms, read the LICENSE file in the repository rather than the badge, and if there is no LICENSE file, ask the maintainer. This is a factual gap, not a legal opinion.
The last push to the repository was on 2026-03-19. That is roughly six months before today, so calling it actively developed would overstate things. The README's own update badge references March 2026 and v3.1, which is consistent with a project that ships in occasional batches rather than continuously.
Upgrade cost is low in absolute terms and hard to measure in practice. Skills are plain text, so a diff tells you exactly what changed in a framework. The friction is that there are no releases to pin and no documented rollback, so a team standardising on these files has to decide its own versioning. The paid Voice Extractor adds a separate purchase and a separate update path on Claw Mart, which the README does not describe. Budget for reading CHANGELOG.md before each pull, and for keeping your own copy of any skill you have modified.
Editorial conclusion
Adopt it if you already run Claude Code and want positioning, content and audit frameworks loaded automatically instead of retyping prompts. Skip it if you need research skills that pull live data without a separate search tool, or if you cannot work in the Claude Code skills folder at all. Before committing, open two or three SKILL.md files, confirm the free set matches the work you actually do, and check whether the research skills you want depend on a paid subscription the README does not spell out.
Frequently asked questions
What are good skills for Claude Code in this repository?
The README groups them by job: strategy and positioning, content and authority, research and intelligence, and conversion and sales. Concrete examples include Positioning Basics, Homepage Audit, LinkedIn Authority Builder and Last 30 Days Research. The README states the skills are designed for Claude Code and could also be used with other agents such as OpenClaw.
What are some effective AI skills for marketing in this repository?
The README lists 23 free skills plus 10 pro, covering positioning, content generation, LinkedIn strategy, cold outreach, case studies and testimonials. Several entries, such as Homepage Audit and Positioning Basics, need only your input and the model, while research skills such as Reddit Insights depend on retrieval the README does not specify.
How do you use Claude for marketing with these skills?
You clone the repository, copy its contents into ~/.claude/skills, and then prompt Claude Code with phrasing that matches a skill's trigger list, for example asking why a landing page is not converting. The README states the agent will then use the matching skill automatically.
What market research skills does ai-marketing-claude-code-skills include?
The README lists Last 30 Days Research, which covers Reddit, X and the web over the past 30 days, Reddit Insights for semantic search over Reddit content, and YouTube Summarizer for transcripts and structured summaries. The README does not name the search backend or any API key these depend on, except for Social Card Generator, which it says has no API dependency.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/brianrwagner-ai-marketing-claude-code-skills)