What problem does it solve? AI-generated text carries recognizable fingerprints—formulaic structures, promotional language, em-dash overuse, vague attributions, and emotion-summarizing narration—that make content feel robotic and untrustworthy. This Skill detects and rewrites those patterns so text reads like it was written by a real person. ## Core Features & Use Cases - Narrative De-AI Layer (fiction): Applies a five-pass method to novels and stories—removing emotion summaries, making characters self-protective instead of self-explanatory, adding cost to plot choices, replacing decorative details with functional ones, and breaking overly tidy paragraph structures. - Language De-AI Layer (general): Detects 24 documented AI writing patterns from Wikipedia's "Signs of AI writing", including inflated significance, -ing shallow analysis, rule-of-three overuse, negative parallelism, AI vocabulary, and Chinese-specific attribute stacking (translation-tone long modifiers before nouns). - Quality Scoring: Rates rewritten text on a 50-point rubric across directness, rhythm, reader trust, authenticity, and concision. - Use Case: Paste an AI-drafted novel chapter or marketing article; the Skill identifies each machine-flavored pattern, rewrites the passages with concrete actions and varied rhythm, and returns the humanized version with a change summary. ## Quick Start Review this article and rewrite it to remove all AI writing traces while keeping the original meaning and tone.
