appautomaton/webmaton
5 skills26 installs90 starsGitHub
Install
npx skills add https://github.com/appautomaton/webmatonSkills in this repo
1Playwright Cliplaywright-cli is a shell interface to Playwright that keeps browser state across commands and exposes page snapshots with stable element refs. A developer uses it to open or attach to browser sessions, click, type and fill forms, inspect console, network and storage, run snippets, capture screenshots or traces, and debug Playwright tests. It prints the Playwright code it ran so interactions can be turned into test code. It is not meant for static HTTP fetches, JSON APIs or simple web search.6installs2Agentic SearchThis skill runs Grok-primary deep web research that returns source-backed, cited answers. A developer uses it when a task needs current web research, high-fidelity page-to-Markdown fetching, site mapping, verbatim quote extraction, or source reranking beyond ordinary search. It ships six scripts (search, fetch, map, extract, rank, get-sources) and persists a session_id so multiple research steps can be composed and revisited.5installs3Chrome Devtools CliWraps the chrome-devtools CLI to automate a debuggable Chrome browser from the terminal using accessibility-tree UID snapshots. A developer uses it to inspect pages, debug runtime behavior, or run Lighthouse and performance traces without setting up an MCP client.5installs4Html To Markdownhtml-to-markdown converts a URL or raw HTML into clean Markdown along with metadata, links, images and quality signals. It wraps nodriver for browser capture of JS-heavy pages and markmaton for the HTML-to-Markdown conversion in one shell pipeline. A developer uses it when an agent or content task needs page text as structured Markdown. It defaults to headless capture but recommends a plain fetch for static articles and server-rendered docs.5installs5Nodriver Browsernodriver-browser is a persistent Chrome or Chromium browser automation skill built on nodriver. It auto-starts a headless or headed Chrome daemon and keeps one tab alive across calls so an agent can navigate, snapshot the DOM, click, type and screenshot in multi-step flows. A developer uses it when a page needs JavaScript rendering, logged-in session continuity, or interaction that WebFetch and search cannot do. It is not meant for static pages, simple searches, JSON APIs or one-off scrapes.5installs