nanaism/yomiyasu

AI生成の日本語を自然な日本語へ推敲するAgent Skill / Agent Skill for Refining AI-Generated Japanese into Natural Japanese

What it solves

It addresses the common issue of "AI-Slop" in Japanese text—where AI-generated content suffers from unnatural metaphors, ambiguous subject-verb relationships, excessive decoration, and a lack of clarity in technical or business documentation.

How it works

It functions as a specialized "skill" for AI coding agents like Claude Code, Cursor, or Codex. It applies seven core transformation principles, such as clarifying subject-verb relationships, removing unnecessary personification, simplifying metaphors, and maintaining consistent tone. Version 2.0.0 introduces a Python-based checking tool (yomiyasu_check.py) that performs linting and diffing to ensure that the AI's revisions do not inadvertently alter the meaning, values, or facts of the original text.

Who it’s for

It is designed for professionals who use AI to draft technical articles, design specifications, PR descriptions, and internal reports, and who need to ensure the output is readable, accurate, and professional without losing the original intent.

Highlights

  • Seven Transformation Principles: Focuses on structural clarity, maintaining the original intent, and removing AI-specific stylistic quirks.
  • Context-Aware Revision: Allows users to specify the document type (e.g., PR, incident report, manual) and the relationship between the writer and reader to adjust tone and formality appropriately.
  • Fact-Checking Support: Includes mechanisms to verify that revisions do not introduce hallucinations or change critical data points by cross-referencing provided materials.
  • Integrated Validation: Bundled with Python scripts to lint and compare the "Before" and "After" versions, providing a record of changes to help users verify the quality of the output.

Related

  • Project
    coji/natural-japanese

    An agent skill that helps AI write more natural, professional Japanese by enforcing writing principles and using morphological analysis to detect and remove common AI-generated cliches.

  • Project
    alchaincyf/huashu-skills

    huashu‑skills is a curated index of 57 Claude‑compatible *Agent Skills* (design, data, video, persona, etc.). 22 lightweight skills live in this repo; the rest are separate repos. Install with a one‑line `git clone` (or copy a sub‑folder), restart the agent, and the skill becomes usable. An included updater checks for newer versions. The collection provides ready‑made, plug‑in functionality for AI agents without writing code.

  • Project
    conorbronsdon/avoid-ai-writing

    A portable writing skill for AI agents that audits and rewrites content to remove formulaic AI writing patterns and "AI-isms" using a tiered vocabulary and pattern catalog.

  • Project
    op7418/Humanizer-zh

    Humanizer‑zh is a Claude skill that rewrites Chinese prose to remove filler and repetitive phrasing while preserving facts, certainty, and the author’s voice. Install via `npx skills add …` or by cloning into `~/.claude/skills/`. Invoke with `/humanizer-zh` on inline text or files. It follows a 31‑point rule set (categories A‑F) and includes a test suite. MIT‑licensed.

  • Project
    harshaneel/humanize

    humanize is an MIT‑licensed repo that ships two static, rule‑based skills—`humanize` (rewrites AI text to sound human) and `ai-check` (produces a forensic AI‑detection report). The skills work with any LLM agent (Claude Code, Codex CLI, ChatGPT desktop, Gemini, etc.) via a simple install script or marketplace command. Benchmarked on 25 diverse inputs, both internal and external perplexity‑based detectors rate the rewritten output as human‑like. The approach is offline, zero‑dependency, but cannot evade learned classifiers like GPTZero.