Brand Voice

This enforces brand voice on everything you write through Claude, from emails to pitch decks to Slack messages. It loads guidelines from session context, a local `.claude/brand-voice-guidelines.md` file, or prompts you to generate them. The implementation distinguishes voice (constant personality attributes) from tone (flexes by content type and audience), which is the right mental model. It validates output against a "We Are / We Are Not" framework and explains its brand choices. Works with companion skills for discovering brand materials and generating guidelines from scratch. The strictness level is configurable, and it surfaces unresolved positioning questions when they affect content. Solid if you're tired of rewriting AI output to sound like your company.