This skill transforms Claude into a sophisticated editorial board for long-form projects such as guidebooks, playbooks, and essays. It employs a unique multi-agent architecture, launching five parallel 'judge' subagents to evaluate content against specific dimensions: practicality, data freshness, language clarity, substance, and completeness. By synthesizing findings into actionable proposals and strictly avoiding edits without user consent, it provides a rigorous feedback loop that eliminates 'LLM-isms' and ensures every sentence serves a clear purpose.
주요 기능
01Support for vision documents and source files to ensure alignment with project goals
02Section-by-section iterative review process for granular control over long documents
0397 GitHub stars
04Evidence-based proposals that quote problematic text with specific rewrite suggestions
05Parallel Multi-Dimensional Critique using five specialized judge subagents
06Automated 'LLM-ism' detection and removal to ensure natural human tone
사용 사례
01Auditing existing documentation for outdated information and stylistic inconsistencies
02Refining technical guidebooks and playbooks for professional executive audiences
03Polishing essays or research papers to remove fluff and improve data citations