# ai-marketing-claude-code-skills: nineteen playbooks a copy command installs > The repository is a directory of skill files that install into a dot folder and load when you mention a related topic, grouped by job rather than by platform. The v3.1 change gave every free skill three execution depths, and the tree carries more directories than the guide lists. **BrianRWagner/ai-marketing-claude-code-skills** — Marketing frameworks that AI actually executes. Designed for Claude Code, but could be used for OpenClaw, etc. - Repository: https://github.com/BrianRWagner/ai-marketing-claude-code-skills - Website: https://www.brianrwagner.com/ - Stars: 427 · Forks: 100 - Language: Shell - License: not declared - Published: 2026-09-10 · Updated: 2026-09-10 · Language: en - Canonical page: https://hysenlabs.com/projects/brianrwagner-ai-marketing-claude-code-skills ## Installation is a copy into a dot directory, and loading is by topic There is no package to install and no manifest. The whole setup is three shell commands: ```bash # Clone the repo git clone https://github.com/BrianRWagner/ai-marketing-claude-code-skills.git # Copy skills to Claude Code's skills folder mkdir -p ~/.claude/skills cp -r ai-marketing-claude-code-skills/* ~/.claude/skills/ ``` The mechanism underneath is an open standard rather than a vendor convention. Agent Skills is described as an open format for packaging expertise as instructions an agent can follow, and the argument for it is stated in a two line contrast: with traditional content you read it, apply it, and forget half of it, while with a skill the agent reads it, applies it, and does so every time. The copy is worth reading carefully, because it copies the repository contents, not a subdirectory. Every skill directory plus the scripts directory, the changelog, and the mode documentation file all land in the same destination. So if you already keep your own skills in that folder, this merges into it rather than namespacing itself, and there is no uninstall instruction in the visible documentation to reverse it. The loading is automatic and topic driven. The claim is that the client will use the skills when you mention related topics, and each skill in the guide lists the phrases that should trigger it, which is a plain-language description rather than a match pattern. ## Three execution depths on every free skill is the whole v3.1 change The repository has a separate file at its top level dedicated to the mode pattern, and the release note for the current version points at it. In March 2026, version 3.1, all nineteen free skills gained support for three execution modes, chosen as a quick, standard, or deep setting, with the stated benefit being fast answers or deep audits from the same skill without extra prompting. That is a more consequential design change than it looks. Without it, a single skill has one cost and one depth, so a user who wants a quick answer and a user who wants a full audit either both overpay or one of them is underserved. Three depths let the same instruction file serve both, which means the trigger phrases stay the same and only the depth argument changes. The phrasing of the release note is also worth noting, because it frames the cost explicitly. The point is not that the deep mode is better but that you can choose, and that choosing does not require rewriting the request to ask for less. The number nineteen refers to the free skills. That matters when reading the repository, because the directory listing at the top level contains more entries than nineteen, which is a gap worth looking at before you assume you have seen everything. ## The catalogue is grouped by job, and each entry declares when it fires The guide organises the skills into five groups, and the grouping is by task rather than by tool, which is the right axis for something that is really a set of prompts. Strategy and positioning holds three. Positioning basics is a core framework for clarifying who a product is for, what it does, and what makes it different. The discoverability audit looks at how a brand appears in AI-powered search and recommendation systems, naming ChatGPT, Perplexity, Claude, and Gemini, and it is triggered by phrases about showing up in those systems. Marketing principles applies named business thinking from a list of practitioners rather than inventing its own, which is a skill whose entire value is whose ideas it cites. Content and authority holds five, and it is the group with the most individual voice in it. LinkedIn authority builder covers positioning, content pillars, formats, and posting rhythm as one system. Content idea generator works from positioning and a stated ideal customer profile. The voice extractor produces a documented voice guide from samples. The de-AI-ify skill removes AI phrasing from text. The social card generator turns one input into platform-specific variants for several networks, and it is the one entry that states no API dependency. Research and intelligence holds four, conversion and sales four, and productivity and operations the remainder. Every entry carries a line of example phrases, which doubles as documentation of when the agent is expected to reach for it and as a way to test whether the skill fires on the wording you actually use. ## Two content skills are about voice, and one of them has a paid twin The voice extractor and the de-AI-ify skill are the pair that deal with how text sounds rather than what it says, and they are the two entries where the project's own claims are most specific. The voice extractor takes samples of someone's writing and produces a documented voice guide, described as being for AI training or ghostwriting, and it triggers on phrases like capturing my voice or write like me. The de-AI-ify skill goes the other direction, removing AI-generated jargon and restoring a human voice, and it carries a provenance claim: it was built from analysing more than a thousand pieces of content, paired AI output against human output. The voice extractor is also the one skill in the list that is not the same artefact for everyone. The entry links to a paid premium version on a separate marketplace at a listed price of nine dollars, alongside the free one. So the free skill and the paid listing are different products with the same name in the catalogue, and a reader comparing them should not assume the paid one is a donation of the free one. That is worth flagging as a pattern rather than a criticism. The rest of the catalogue is free with no upsell, and the project is otherwise consistent about that. One entry having a paid version is a business decision, and the honest thing is that the guide does not explain what the paid one adds. ## The research skills reach outside your machine, and one states its time saving Four skills sit in the research group and they behave differently from the rest of the catalogue, because they depend on sources you do not control. The last 30 days skill searches across Reddit, X, and the web for a topic from the previous month, and it makes a specific claim about the saving: seven minutes instead of two hours of manual research, with current trends and real community sentiment as the output. That is the only performance figure in the guide, and it is a comparison against a manual baseline rather than a benchmark. The Reddit insights skill is described as using semantic AI search rather than keyword matching, and its stated purposes are finding user pain points, discovering niche markets, and validating a business idea with real feedback. The YouTube summariser fetches video transcripts and produces structured summaries with key insights and metadata, and it notes that full transcript delivery is supported. The LinkedIn profile optimiser is the fourth, and it scores each section of a profile, rewrites the headline and about copy, and includes a visibility checklist for AI search. The practical distinction is between the skills that operate on text you paste and the skills that go and fetch. A voice guide is derived from something you supply; a trend report is derived from a platform's API behaviour on the day you run it. The first is reproducible, and the second is not, which matters if you are building anything on top of the output. ## The directory listing holds more entries than the guide describes The guide says nineteen free skills, and the top-level listing contains more than that. Alongside the directories the guide walks through, there is a directory whose name suggests a different mode entirely, a scripts directory, the changelog, the mode pattern file, and a file with a date in its name that reads like a plan document. That gap is the most useful thing to notice before installing, because the copy command copies the whole repository contents into your skills folder. Whatever is in those extra directories ends up there too. The conversion and sales group is also fuller than the marketing framing suggests. Alongside a homepage and landing page audit, a cold outreach sequence builder, a case study builder, and a testimonial collector, there is a vault cleanup auditor and a tweet draft reviewer in the listing. The names read like an operations toolkit for someone who runs their own marketing rather than a set of frameworks for a team, which is consistent with the rest of the catalogue being written in the first person. The project publishes no GitHub releases, so there is no tag to pin to and no upgrade path beyond copying the repository again. The last push to the repository is dated 2026-03-19, which is the date to weigh when deciding whether to depend on it. One other detail for anyone assessing reuse: the guide carries a licence badge, but the repository's licence field is not filled in, so the terms on the files themselves are worth checking directly rather than inferring from the badge. ## Conclusion This suits someone who writes marketing copy or does positioning work and wants a repeatable structure instead of improvising a prompt each time, and who is comfortable dropping files into a dot directory. It is a poor fit if you want a tool that does something, since most of these are instruction files rather than programs, and a poor fit if you need a maintained library, since the last push to the repository was 2026-03-19 and there are no releases. Before you install, check three things: that the copy command will not clobber skills you already have in the same folder, because it copies the whole repository contents in; which depth you want by default, since the three execution modes change both the effort and the length of what you get back; and whether any of the skills reach outside your machine, since the research ones search Reddit, X, and the web and the video one fetches transcripts. One skill also has a paid version listed separately, so the free and paid versions of that one are not the same artefact. ## FAQ ### How do I install the ai-marketing Claude Code skills? Clone the repository, create the skills folder if it does not exist, and copy the repository contents into it with mkdir and cp. There is no package to install and no manifest, and the client then uses the skills automatically when you mention a related topic. ### What execution modes do the ai-marketing skills support? Every free skill supports a quick, a standard, and a deep mode. That arrived in version 3.1 in March 2026 and is documented in a separate pattern file at the top of the repository, so the same skill can give a fast answer or a deep audit without extra prompting. ### What are the skill categories in this Claude Code marketing repository? Five groups: strategy and positioning, content and authority, research and intelligence, conversion and sales, and productivity and operations. The grouping is by job rather than by tool, and every entry lists example phrases that should trigger it. ### Do any of these marketing skills need an API key or send my data somewhere? The social card generator is the one entry that states no API dependency. The research skills do reach outside your machine: one searches Reddit, X, and the web, another uses semantic search on Reddit, and the video summariser fetches YouTube transcripts, so their output depends on those platforms at the moment you run them. ### Is the voice-extractor skill in this repository free? The free version is in the repository, and the guide also links to a paid premium version of the same skill on a separate marketplace at a listed price of nine dollars. The two are separate listings, and the guide does not describe what the paid one adds. ## Sources - [BrianRWagner/ai-marketing-claude-code-skills on GitHub](https://github.com/BrianRWagner/ai-marketing-claude-code-skills) - [Issues](https://github.com/BrianRWagner/ai-marketing-claude-code-skills/issues) - [Project website](https://www.brianrwagner.com/) - [README](https://github.com/BrianRWagner/ai-marketing-claude-code-skills/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/brianrwagner-ai-marketing-claude-code-skills