Guideline Generation

Takes your brand docs, sales call transcripts, or discovery reports and turns them into structured voice guidelines that LLMs can actually follow. It builds a "We Are / We Are Not" table, tone-by-context matrix, and assigns confidence scores to each section so you know what's solid versus what needs team input. The quality check agent verifies you get at least 3 voice attributes with evidence before it saves anything. What's smart is how it handles ambiguity: every open question includes an agent recommendation, so you're confirming or overriding instead of starting from scratch. Saves to `.claude/brand-voice-guidelines.md` and archives your previous version automatically.

Install to Claude Code

npx -y skills add anthropics/knowledge-work-plugins --skill guideline-generation --agent claude-code