ai-marketing-claude vs designer-skills
15 marketing skills for Claude Code — /market audit runs 5 parallel agents scoring a site across 6 dimensions; copy, email sequences, ad creative, competitor intel and client-ready PDF reports. — versus — 239 design skills, 88 commands and 33 plugins for Claude Code and Gemini CLI — research, design systems, UI, interaction and delivery, written for agents to actually execute.
Same shape, different profession: both are domain-expertise skill packs that turn Claude Code into a practitioner — designer-skills for research→design→delivery, this one for audits→copy→campaigns→client reports.
| ai-marketing-claude | designer-skills | |
|---|---|---|
| Stars | 2.1k | 1.8k |
| Forks | 635 | 307 |
| Language | Python | Markdown |
| License | MIT | MIT |
| Last activity | 4 months ago | 1 months ago |
| Topics | skills | skills |
| Curated connections | 3 | 4 |
ai-marketing-claude — the curator's take
The 'sell marketing services with Claude Code' starter kit: one command fans out five subagents that score a site on content, conversion, SEO, positioning, brand and growth, and the deliverables (weighted scores, before/after copy, client proposals, PDF reports) are shaped for agency work, not just self-serve tinkering. Templates for welcome/nurture/launch sequences included. Know what it is: prompts and scoring rubrics, not integrations — nothing talks to ad accounts or analytics APIs (claude-ads is the operations side of this coin). The scoring is LLM-judged, so treat numbers as directional. Repo hasn't moved since March and the README funnels to a paid Skool community — the MIT code stands alone fine.
designer-skills — the curator's take
The largest coherent skill pack for design work: five collections covering research→systems→UI→interaction→delivery, installable straight from Claude Code's plugin marketplace with genuinely non-technical instructions. If you're a designer adopting agents (or an engineer faking design taste), this is the fastest on-ramp. NOT a design tool itself — it's prompts-as-skills, so output quality tracks the underlying model, and 239 skills means uneven depth: audit the ones you'll actually use rather than installing all five collections on day one.