ai-text-humanizer

To detect AI generated Japanese text patterns, this Skill applies regex-based detection across six categories to compute an AI-smell score from 0 to 100. It identifies specific unnatural writing patterns and returns per-pattern evidence with line-level recommendations.

The best way to rewrite LLM output for natural Japanese is using rule-driven rewriting techniques that break balance, objectivity, and rigid logic. This Skill applies these humanization methods to produce a before-and-after improvement workflow with structured reporting.

Regex pattern detection for AI tone scoring is optimized exclusively for Japanese text. While English text processing relies on Claude's internal reference guidance, the script-based regex detection and pattern scoring mechanisms specifically target Japanese writing structures.

To check AI tone in Japanese business documents and emails, the Skill evaluates text against six AI-specific writing patterns. It generates a 0 to 100 quantification score and provides actionable improvement recommendations tailored for editing contexts like blog drafts and chats.

Yes, this AI tone detector supports markdown cleanup for blog drafts alongside its core scoring functionality. It evaluates the cleaned text content to identify AI-like patterns and applies humanization rewrites to ensure the final draft reads naturally.

The limitation of rule-driven humanization is its reliance on predefined regex patterns and stored reference guidance, meaning it may miss nuanced contextual subtleties. Detection is strictly optimized for Japanese syntax, making it less effective for English or non-standard text structures.