If you've ever asked an LLM for character names and gotten three variations of Chen, Patel, and Maya, this skill exists to break that median clustering. It uses external entropy through curated cultural name lists and phoneme presets instead of letting the model pick from its statistical defaults. You get scripts that randomize from deep cultural pools, track your cast for collision risks like Mark/Mike/Michael, and generate phonologically consistent fantasy names that don't sound like alphabet soup. The core insight is simple: LLMs can't help defaulting to training data medians, so the only fix is genuine randomization from outside the model. Built by jwynia for fiction writers tired of algorithmically obvious character rosters.
npx skills add https://github.com/jwynia/agent-skills --skill character-naming