pencil20388-eng

pencil20388-eng / stop-slop-zh

Public

🚫 Stop AI slop in Chinese writing. 干掉中文 AI 味 — banned words, punctuation rules, structure constraints & 4-layer QA. Works with Claude Code, Cursor, Codex CLI.

ai-writing anti-ai-slop chinese chinese-nlp claude-code
14
0
89% credibility
Found May 30, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A writing improvement tool that teaches AI assistants to remove robotic patterns from Chinese text, replacing them with natural conversational language and running automatic quality checks.

How It Works

1
😤 Tired of robotic AI Chinese?

You notice your AI assistant writes Chinese that sounds like a textbook or news broadcast instead of a real person.

2
🔍 Find stop-slop-zh

You discover a free tool that teaches your AI to write natural, conversational Chinese.

3
⚙️ Set it up once

You add the writing rules to your AI assistant with a simple one-time setup.

4
✍️ Write or check content naturally

You ask your AI to write articles or check existing text using everyday language.

5
📋 See quality check results

Your AI runs four quality checks and shows you a detailed report on what was fixed.

🎉 Content sounds authentically human

Your Chinese text now reads like a knowledgeable friend having a real conversation, not a machine outputting information.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is stop-slop-zh?

Stop-slop-zh is a prompt-based skill that strips the "AI smell" out of Chinese writing. It works by feeding your AI tool a set of hard rules: banned filler phrases like "综上所述" and "值得注意的是," forbidden punctuation patterns like colons and em-dashes, and structural constraints that kill the numbered-list style that makes AI prose feel robotic. After rewriting, it runs a four-layer quality check and spits out a QA report. Compatible with Claude Code, Cursor, Codex CLI, and any agent that reads SKILL.md files.

Why is it gaining traction?

Chinese AI writing has its own distinct tells, completely different from English patterns. Most tools ignore this entirely. This project fills that gap with a curated rule set built from real content operations work, plus a library of 50+ colloquial replacements ("坦率的讲," "踩过坑," "太离谱了") that make prose sound human. The four-layer QA system is the hook: hard rules first, then style, content, and a "活人感" check that catches the uncanny valley residue. It's the Chinese counterpart to the popular English stop-slop project.

Who should use this?

Content teams producing Chinese marketing copy, documentation, or social posts who are tired of AI output that reads like a textbook. Developers integrating AI into Chinese content workflows will find it useful for post-processing. If you've ever asked an AI to write something in Chinese and cringed at the result, this is for you.

Verdict

With 14 stars, this is early-stage but the concept is solid and the banned-word list alone saves hours of manual editing. The 0.9% credibility score reflects its niche audience, not its utility. Try it if you write Chinese with AI regularly -- just don't expect polished docs or community support yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.