xiaomoBoy

A toolkit of Claude Code skills for long-form content creators: research, score, rewrite, and publish.

17
0
85% credibility
Found May 25, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

This is a toolkit of AI assistant skills designed for content creators who write long-form articles and publish them across multiple platforms. It provides 5 specialized helpers: one collects research from YouTube and the web into organized summaries, one scores articles against a viral-content quality rubric and suggests improvements, one tunes the scoring system itself using your own examples, one fixes image links so they work on any platform, and one helps download videos or audio when you need source material. The project is well-documented in both English and Chinese, integrates with popular AI coding tools like Claude Code, and includes step-by-step setup guides for different computer systems.

How It Works

1
💡 Discover a Writing Helper

You hear about a tool that helps you research, score, and publish articles to multiple platforms without doing each step manually.

2
🧩 Add It to Your Assistant

You connect the toolkit to your AI writing assistant with a simple command, and suddenly you have 5 new abilities ready to use.

3
🔍 Gather Research Material

When you want to write about a topic, your assistant collects YouTube videos and web articles for you, organizes everything, and summarizes the key points.

4
📝 Write and Refine Your Article

You write your article naturally, then ask your assistant to score it against what makes content popular. It tells you which parts could be stronger.

5
Two Paths to Improve
✍️
Rewrite the article

Your assistant helps you rewrite the weak parts, then scores again to show improvement

🎯
Calibrate the rubric

You provide examples of good and bad articles, and adjust what the scoring system considers important

6
🖼️ Fix Image Paths

Before publishing, the tool automatically fixes any image links so they will display correctly on all platforms.

🚀 Publish Everywhere at Once

Your finished article goes out as drafts to all your favorite platforms — Chinese websites, blogging sites, and more — ready for you to review and publish.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 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 claude-writing-skills?

This is a Python toolkit of five Claude Code skills designed for long-form content creators who want to automate their research-to-publish workflow. The skills handle discrete tasks: gathering sources from YouTube and the web, scoring articles against a viral-content rubric, iteratively rewriting to improve scores, calibrating the scoring system itself, and downloading media with yt-dlp. Each skill is self-contained and composable, so you can pick just the pieces that fit your pipeline. The project integrates with real CLI tools like notebooklm-mcp-cli, codex, wechatsync, and yt-dlp rather than reinventing functionality in prompts alone.

Why is it gaining traction?

The hook is the article optimizer loop. Instead of manually revising based on gut feel, you get a 9-dimension score, target the weakest areas, and iterate with measurable progress. The companion score optimizer lets you tune the rubric itself against your own labeled samples, which is genuinely useful for anyone publishing to algorithm-driven platforms. The documentation is thorough across English and Chinese, with per-OS install guides and integration recipes for Cursor, Aider, and Codex CLI. The design principle of "one skill, one job" keeps boundaries clean and makes debugging predictable.

Who should use this?

Writers and developers who use Claude Code for content production will get the most value. If you publish to Chinese platforms like 知乎 or 掘金, the wechatsync integration saves real friction. The scoring skills appeal to creators optimizing for viral reach who want data-driven revision cycles. It's less useful if you need an end-to-end writing assistant or don't use Claude Code.

Verdict

At 17 stars, this is early-stage but actively maintained with solid documentation. The credibility score of 0.85 reflects a well-organized repo with clear use cases, though the small community means limited real-world stress testing. Worth trying if your workflow maps to the five skills, but treat it as a toolkit to adopt piecemeal rather than a monolithic solution.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.