Brian Wagner has published an open-source repository of 24 marketing frameworks packaged as executable skills for AI coding agents. The collection is optimized for Claude Code but tested and functional on OpenClaw, GitHub Copilot, and Cursor, built on the open Agent Skills standard.
How the Skills Work
Each skill is a self-contained SKILL.md file with defined triggers, execution steps, and output formats. When a user types a natural language command like “help me with positioning” or “audit my homepage,” the agent loads the corresponding skill and follows a structured framework rather than generating advice from scratch.
The March 2026 v3.1 update added three execution modes across all 19 free skills: quick (fast answers), standard (balanced analysis), and deep (comprehensive audit). According to BrightCoding’s review, the Last 30 Days Research skill crawls Reddit, X, and the web to deliver community sentiment reports, while the AI Discoverability Audit targets how brands appear in ChatGPT, Perplexity, Claude, and Gemini results.
The 24 skills span five categories: positioning, content creation, research, conversion optimization, and productivity. Nineteen are free and five are premium, with the premium Voice Extractor skill priced at $9 on Claw Mart.
Cross-Platform Installation
Installation follows a copy-to-directory pattern native to each platform. Claude Code users clone the repo and copy skills to ~/.claude/skills/. OpenClaw users target ~/.openclaw/skills/. Copilot and Cursor have their own skill directories. The GitHub repository includes compatibility badges and tested configurations for each platform.
The interoperability comes from the Agent Skills standard, which defines a specification for SKILL.md files with sections for triggers, parameters, execution logic, and output formats. Any agent platform that can parse and follow markdown-structured instructions can execute these skills.
Ecosystem Signal
The repository is one data point in a broader pattern: skill libraries are becoming the packaging format for reusable agent capabilities. OpenClaw has its own built-in skill system. Anthropic’s Claude Code supports skill loading natively. The question is whether a shared standard like Agent Skills becomes the common interchange format, or whether each platform develops its own incompatible skill ecosystem.
For builders running agents across multiple platforms, cross-compatible skill libraries reduce the cost of switching or running heterogeneous agent setups. Whether marketing is the right beachhead for proving skill interoperability depends on adoption, but the repo demonstrates that the file format works across at least four agent platforms today.