Humanly AI Writing Refiner
Removes AI-generated writing patterns and robotic 'slop' to create natural, human-sounding content in multiple languages.
Humanly is a specialized Claude Code skill designed to strip away the predictable artifacts and formulaic structures often found in AI-generated text. By utilizing a multi-tiered analysis system based on industry-standard AI writing indicators, it helps solopreneurs and developers transform robotic outputs into authentic, engaging prose. The skill features three distinct modes: a pre-writing reference to internalize humanizing rules, a comprehensive review pipeline for post-writing audits, and a detection mode for flagging 'AI tells' without altering the source material.
주요 기능
01Comprehensive 50-point quality scoring system based on rhythm, authenticity, and directness
0256 GitHub stars
03Bilingual support with optimized rulesets for both English and Chinese (去 AI 味)
04Second-pass audit logic to ensure surviving AI tells are identified and corrected
05Context-aware profiles for social media, technical blogs, investor emails, and documentation
06Multi-tier detection for credibility-killers (P0), obvious AI smell (P1), and stylistic polish (P2)
사용 사례
01Auditing professional emails and social copy to remove significance inflation and chatbot artifacts
02Refining AI-generated blog posts and documentation to sound like a 'smart friend talking'
03Pre-briefing Claude with anti-slop rules before drafting public-facing communication