Content Skill Graph

Most AI content agents produce copies, not posts. Hand them a flat brand voice file and ask for five LinkedIn posts — you get five versions of the same idea with synonyms swapped. The format is wrong before the writing even starts.

Content Skill Graph fixes the architecture. Instead of one flat file your AI reads once, you get 11 interconnected markdown nodes it traverses — voice node, platform node, engine node, audience node. Every traversal path produces a genuinely different output. Same topic, three platforms, three native posts.

What's included: a complete 11-node wikilink graph with copy-paste templates for every node — brand voice, anti-patterns, tone calibration, LinkedIn rules, X rules, hook formulas, post structures, repurpose logic, and audience profiles. Plus the 7 Voice Checks QA system that ships zero flat-file content through your pipeline.

Who it's for: marketers and founders running AI content agents who are tired of every post sounding like the same brief repackaged. Works with Claude, GPT-4, Scribe, or any LLM-based content agent that can follow wikilinks.

Your AI is already a graph traverser. The question is whether you gave it a graph worth traversing.