Detects and refines AI-generated patterns in academic drafts to restore authentic human voice and improve prose quality.
Writing De-AI is a specialized skill designed to help researchers identify and mitigate telltale signs of LLM-generated text in academic manuscripts. By analyzing text across ten distinct categories—such as hedging overuse, formulaic transitions, and uniform sentence length—the skill generates a comprehensive score and a detailed report. It identifies the most significant 'AI-likeness' issues and provides context-aware rewrite suggestions that preserve technical accuracy and citations while enhancing the authorial voice, making it an essential final check before journal submission.
주요 기능
01Preservation of technical claims and citation integrity
02Automated scanning of LaTeX (.tex) and Markdown (.md) paper drafts
03Targeted rewrite suggestions for the top 5 worst stylistic offenders
04Context-aware identification of generic fillers and formulaic transitions
051 GitHub stars
06Comprehensive 50-point scoring system across 10 detection categories
07Detailed reporting including sentence-level issue logging and scoring
사용 사례
01Identifying robotic or monotonous sentence structures in long-form technical content
02Humanizing the introduction and discussion sections of academic manuscripts
03Polishing AI-assisted research paper drafts before peer review
