OpenCLI Web Automation Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection. OpenCLI turns any website into a command line interface by reusing Chrome's logged in browser session. It supports 19 sites and 80+ commands out of the box, and lets you add new adapters via TypeScript or YAML dropped into the clis/ folder. Installation Prerequisites Node.js = 18.0.0 Chrome browser running and logged into the target site [Playwright MCP Bridge](https://chromewebstore.google.com/detail/playwright mcp bridge/mmlmfjhmonkocbjadbfplnigmagldckm) extension installed in Chrome Install from Source (Development) Environment Configuration MCP client config (Claude/Cursor/Codex ~/.config/ /config.json ): Key CLI Commands Discovery & Registry Running Built in Commands Output Formats All commands support format / f : AI Agent Workflow (Creating New Commands) Explore artifacts are saved to .opencli/explore/<site / : manifest.json — site metadata endpoints.json — discovered API endpoints capabilities.json — inferred command capabilities auth.json — authentication strategy Adding a New Adapter Option 1: YAML Declarative Adapter Drop a .yaml file into clis/ — auto registered on next run: Option 2: TypeScript Adapter Common Patterns Pattern: Authenticated API Extraction (Cookie Injection) Pattern: Header Token Extraction Pattern: DOM Scraping with Wait Pattern: Pagination Maintenance Commands Testing Troubleshooting Symptom Fix Failed to connect to Playwright MCP Bridge Ensure extension is enabled in Chrome; restart Chrome after install Empty data / Unauthorized Open Chrome, navigate to the site, log in or refresh the page Node API errors Upgrade to Node.js = 18 Token not found Run opencli setup or opencli doctor fix Stale login session Visit the target site in Chrome and interact with it to prove human presence Debug Verbose Mode Project Structure (for Adapter Authors)

opencli-web-automation