Most content workflows still involve stitching together separate tools: a model for text, another for images, manual formatting, then a browser for publishing. This repo centralizes those pieces into small, composable skills you can call from Claude Code — turning markdown and prompts into publish-ready posts, visuals, and automation scripts.
What Sets It Apart
- Modular skill surface: individual directories like
skills/baoyu-*are treated as independent plugins. This lets teams adopt only the pieces they need (image generation, slide decks, social posting) without pulling the whole repo. - Provider-agnostic generation: image and text-generation skills include adapters for many backends (OpenAI/Azure/OpenRouter/Google/Replicate/MiniMax/Seedream/etc.), so you can swap providers by changing environment keys rather than code.
- End-to-end publishing glue: includes browser-driven CDP flows and API paths for publishing to platforms (X/WeChat/Weibo) and tooling to publish skills to ClawHub/OpenClaw, reducing manual copy-paste and UI interaction.
- Practical creator focus: features like YouTube transcript extraction, markdown→HTML, infographic/cover/slide image generators, and compression/format utilities show the project targets content creators who want repeatable, scripted publishing pipelines.
Who It's For & Trade-offs
Great fit if you are a developer or content team that: wants to automate article-to-image or article-to-slide workflows; already uses Claude Code; and needs multi-provider image/text generation with reproducible CLI flows. It accelerates publishing and prototyping by packaging common tasks as invokable skills.
Look elsewhere if you need a standalone hosted SaaS, a non-Node runtime, or a managed LLM inference service — this repo is a developer-focused, local/CLI-first toolkit that expects you to configure API keys and run skills in your environment. Some skills rely on unofficial browser automation (CDP) which carries maintenance risk if target sites change.
Where It Fits
Think of this project as a curated marketplace of automation building blocks for Claude Code: higher-level than single-model wrappers but lighter-weight than a full platform. It complements LLM/image provider SDKs by focusing on content-generation workflows and publishing connectors rather than raw model research or training.
