codex-article

Researches and writes evidence-driven Chinese AI articles with fact-checking and title generation.

17.5k|2.5k|Updated Nov 8, 2018
One-click install
npx skills add https://github.com/itwanger/toBeBetterJavaer --skill codex-article
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-article
Source: https://github.com/itwanger/toBeBetterJavaer/tree/main/.agents/skills/codex-article
Command: npx skills add https://github.com/itwanger/toBeBetterJavaer --skill codex-article

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Writing credible Chinese AI industry articles requires verifying facts against primary sources, structuring arguments around evidence, and maintaining consistent style—this Skill automates that entire research-to-publication workflow.

Core Features & Use Cases

  • Evidence-First Research: Verifies claims against official announcements, source code, and reproducible tests before writing, separating confirmed facts from vendor claims and unknowns.
  • Six Article Structures: Provides proven templates for model reviews, price-war commentary, outage analysis, opinion pieces, methodology breakdowns, and security incident retrospectives.
  • Automated Quality Checks: Runs a Python script validating minimum length, greeting format, frontmatter fields, screenshot placeholders, and section numbering.
  • Use Case: Given a working title and links about a new model release, the Skill researches primary sources, writes a 4000+ character article in the author's voice, inserts screenshot placeholders, and generates 5 candidate titles.

Quick Start

Use the codex-article skill to research this AI news material and write a complete article with five candidate titles.

Frequently Asked Questions about codex-article

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a fact-checked Chinese AI article with AI assistance?

Provide a working title and source materials, then the workflow verifies claims against official announcements, source code, and reproducible tests before drafting. It produces a 4000+ character Markdown article with evidence boundaries and five candidate titles.

What article structures does this writing workflow support?

Six structures are supported: single-model reviews, price-war commentary, outage and account governance analysis, industry opinion pieces, official methodology breakdowns, and security incident retrospectives. Each follows a specific evidence-ordering pattern.

How does the article validation script work?

Run check_article.py with the article path and --min-chars 4000 to verify Chinese character count, the fixed greeting, and quote style. Add --require-screenshots to also validate frontmatter fields, screenshot placeholder syntax, and per-section image coverage.

Can I use this workflow for non-AI topics or tutorials?

Installation tutorials and interview articles do not follow the six structures but still enforce evidence discipline, short paragraphs, and fact boundaries. The workflow is optimized for AI industry content including models, agents, RAG, and product news.

What are the limitations of the automated article checks?

The script only performs mechanical checks like length, formatting, and placeholder syntax. Humans must still verify that claims trace to primary sources, comparisons use consistent benchmarks, and screenshots sit near the facts they prove.