Eze

You can run searches, scrape web pages, operate logged-in websites, and read dynamically rendered content through a real Chrome browser via CDP. Additional skills collect and summarize RSS feeds into daily briefs, generate images from text prompts with Doubao Seedream 4.0, and sustain persona-aware conversations with persistent user notes.

They target developers and power users who need real-browser internet access, including social media scraping on platforms like Xiaohongshu, Weibo, and Twitter. They also suit content curators publishing daily news digests and users wanting persistent, context-aware conversational partners with remembered notes and personas.

Web-access routes all联网 operations—search, page fetching, login-required actions, and dynamic page rendering—through a real Chrome browser controlled via the Chrome DevTools Protocol. It is triggered whenever a task requires an authentic browser environment, including social media scraping and interactive web page operation.

The web-access skill is released under the MIT license and is free to use and modify. Its source code is publicly available at github.com/eze-is/web-access, currently at version 2.5.4, maintained by author 一泽Eze.

Web-access requires a Chrome browser instance controllable via CDP for rendering and login sessions. The seedream-image-generator requires access to the Doubao Seedream 4.0 API, while daily-news depends on RSS feed sources and web pages as input for its summarization and publishing pipeline.