The llm-writing skill helps Claude transcend 'default LLM behaviors' that often result in generic or repetitive prose. It works by identifying and resisting common AI tropes, such as filling structure without purpose, stating conclusions without evidence, and using unnecessary conversational transitions like 'Let's break this down.' By eliminating 'conversational bleed,' this skill ensures that generated artifacts—whether they are technical documents or creative pieces—are written specifically for the end reader rather than as a direct response to a chat prompt.
主な機能
01Eliminates conversational bleed to ensure documents are standalone and professional.
02Integrates with intent-modeling to anchor content to specific goals.
03Encourages evidence-based explanations over generic structural summaries.
04Filters out repetitive transitions and redundant sentence structures.
05246 GitHub stars
06Identifies and removes common AI writing tropes and robotic patterns.
ユースケース
01Drafting blog posts and articles that feel human-authored and reader-centric.
02Polishing technical documentation to remove conversational filler and AI-speak.
03Generating standalone reports that do not reference the prompt or chat context.