Retrieve structured browser automation action manuals by ID and keyword.
npx skills add https://github.com/Mte90/dotfiles --skill core-actionbook
Please help me install this Agent Skill. Skill: core-actionbook Source: https://github.com/Mte90/dotfiles/tree/main/.config/opencode/skills/core-actionbook Command: npx skills add https://github.com/Mte90/dotfiles --skill core-actionbook
This Skill provides pre-computed action manuals for browser automation, enabling agents to interact with web pages more efficiently by using structured page information and element selectors instead of parsing raw HTML.
search_actions to find a specific action for logging into a website, then use get_action_by_id to retrieve the exact selectors for the username and password fields, and finally execute these actions to log in.Use the core-actionbook skill to search for actions related to logging into a website.
High-intent search queries and answers about installing and using this skill.
Pre-computed element selectors enable deterministic browser automation by providing agents with exact CSS and XPath selectors, bypassing the need to parse raw HTML. You retrieve these structured action manuals via ID to execute precise web page interactions like logging into a website.
An action book is a structured repository of browser automation manuals. It provides agents with pre-computed DOM structures and CSS or XPath selectors, allowing efficient web scraping and interaction without parsing entire HTML documents.
You retrieve specific CSS selectors by searching for actions with a keyword, then using the returned ID to fetch the detailed action manual. This manual contains the exact DOM structure and CSS or XPath selectors needed for deterministic task execution.
Yes, you can use MCP tools to interact with web pages without parsing raw HTML. This approach provides structured page information and element selectors directly, enabling faster and more reliable web automation by eliminating HTML parsing overhead.
Web automation agents fail when attempting to parse raw HTML for dynamic elements. Using pre-computed action manuals with structured DOM details and deterministic CSS or XPath selectors provides reliable element targeting and prevents parsing failures.
mikeyobrienRetrieve device GPS or network location and output JSON coordinates via Termux.
zinohomeAutomate Chromium-based browser interactions including screenshots, form completion, and JavaScript execution.
theSJaceAutomate browser interactions using verified action manuals and a CLI-driven workflow.
PhucMPhamAutomate browser tasks with Puppeteer scripts for testing and data extraction.