yanhua1010/self-media-content-workflow

What it solves

It streamlines the entire lifecycle of self-media content production, transforming vague ideas into platform-specific posts. It solves the problem of maintaining consistency across different social platforms while ensuring that content is tailored to each platform's native style, avoiding the "one-size-fits-all" approach to posting.

How it works

The project provides a modular set of "Agent Skills" that can be integrated into AI agents (like Claude Code or Cursor). It uses a central controller to route tasks across specialized modules:

• Content Brief & Strategy: Clarifies goals, audience, and creates content calendars.
• Trend Radar: Tracks hot topics and analyzes competitors.
• Platform Copywriting: Generates native text for X, Xiaohongshu, and WeChat, including a visual style library for images.
• Short Video: Plans hooks, scripts, and storyboards.
• Analytics & Delivery: Handles data review, archiving, and publishing to platforms like WeChat.

To prevent AI hallucinations and maintain quality, it implements a "human-in-the-loop" system with five mandatory confirmation points (direction, platform, title, final draft, and publishing authorization).

Who it’s for

Content creators, social media managers, and self-media operators who want to automate the research and drafting process while maintaining strict editorial control and platform-specific optimization.

Highlights

• Tool Agnostic: Does not bind to specific models or browsers; it discovers available capabilities at runtime.
• Platform Native: Shares a single set of facts across platforms but rewrites titles and structures for each.
• Evidence-First: Requires sources for key numbers and prohibits the fabrication of data or experiences.
• Safety-First: Includes strict boundaries against automatic mass-posting, scraping via main accounts, or storing sensitive credentials.