What problem does it solve? AI-generated or AI-assisted text often reads as templated, vague, overly balanced, or emotionally hollow, which undermines credibility with readers. This Skill diagnoses where the "AI flavor" comes from and rewrites text so it sounds like a specific person writing for a specific context, without changing facts, numbers, sources, or meaning. ## Core Features & Use Cases - Six-layer diagnosis: Locates problems at the viewpoint, structure, expression, material, narrative, emotion, or context-mismatch layer instead of just swapping words. - Context-aware rewriting: Adapts output by genre (reports, emails, film reviews, personal letters), platform (Xiaohongshu, WeChat, Zhihu, Bilibili), audience, industry, and publicity level, with a built-in scoring gate before delivery. - Fidelity guardrails: Never fabricates experiences, data, quotes, or testimonials; preserves numbers, citations, and risk disclaimers; refuses AI-detection evasion requests. - Use Case: A user pastes a work report full of phrases like "continuously advanced" and "optimized processes" and says it sounds too AI. The Skill rewrites it with concrete actions, blockers, and next steps while keeping every fact intact. ## Quick Start Rewrite this text so it sounds less like AI and more like a real person, keeping all facts and numbers unchanged.
