dnh33/tasteful-llm
Reduces AI slop in creative and strategic output. Substitution Test gate, per-model profiles for Claude, GPT and DeepSeek, and a 52-pattern anti-slop catalog in six categories.
Use when reviewing existing text for AI slop: landing pages, README prose, pitch decks, marketing copy, product descriptions. Use when the user says 'audit this copy', 'check for slop', 'does this sound like AI', 'tighten this up', 'review this for quality', or 'is this too generic'.
Run a guided taste calibration session that builds a foundational taste profile. Use when the user says '/tasteful-llm:taste-setup', 'taste setup', 'set up my taste profile', 'configure tasteful', 'create taste profile', or when no taste profile exists and the user wants to establish one.
Use when producing prose where word choice shapes perception: marketing copy, product descriptions, taglines, hero sections, pitch text, README narrative sections, naming (brands, products, features), positioning statements, design rationale. Do not use for code, data analysis, debugging, technical docs, commit messages, or structured output like JSON/YAML.