OpenCLI
Turns any website into a CLI command and lets an agent drive the user's already logged-in browser via Playwright
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Acts through the user's real logged-in browser and can take actions under their accounts
- Writes and runs adapters that decide themselves how to interact with a site
Install
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/opencliDetects the agents on your machine, checks the risk and pins the version.
Other ways to install
Install the tool
npm install -g @jackwener/opencliInstall the CLI with npm install -g @jackwener/opencli, install the Browser Bridge extension from the Chrome Web Store, then run opencli doctor to check the connection. Install the agent skills with npx skills add jackwener/opencli.
Other ways from the author
npm install -g @jackwener/opencliAfter installing, run opencli doctor to verify the browser connection.
This is third-party code. Review the repository files before installing.
What it does
OpenCLI ships ready-made adapters for dozens of sites such as Bilibili, Zhihu, Reddit, and Hacker News, which an agent calls like ordinary terminal commands. Separate skills let an agent drive any site directly: navigate pages, click, fill forms, and read content through the user's already logged-in Chrome profile, without a manual login. The opencli-adapter-author skill walks the agent through writing a new adapter for an unfamiliar site, from page reconnaissance to a finished typed command. Beyond websites, the tool also works as a hub for local CLIs such as gh, docker, or discord and can drive commands for desktop Electron apps.
Who it is for. For people who want an agent to browse sites under the user's own account and pull or fill in data there.
Good fit when
- You want an agent to read or gather data from a site that requires a login
- You want a reliable command for a site instead of driving a browser from scratch each time
- You want to add support for a new site quickly, from a template
Not a fit when
- You do not need access to a personally logged-in browser and a public API is enough
- Chrome with the bridge extension cannot be installed
Example request
Go into my Reddit account through the browser and collect the titles of today's top 5 postsLimitations
Requires Node.js 20.18.1 or newer and the Browser Bridge extension for Chrome with a local daemon. The agent gets access to the user's real logged-in profile, so commands should be reviewed before running on sensitive sites.
How to disable. Remove the global package with npm uninstall -g @jackwener/opencli and disable the Browser Bridge extension in Chrome.
Security check
- Acts through the user's real logged-in browser and can take actions under their accounts
- Writes and runs adapters that decide themselves how to interact with a site
README in short
The README describes three usage modes: ready adapters for known sites, the opencli-browser skill for driving any site through the agent's logged-in Chrome, and a hub for local CLIs and desktop apps. Setup is npm install -g, then the Browser Bridge extension from the Chrome Web Store, verified with opencli doctor. For agents like Claude Code and Cursor, skills install with npx skills add, and individual skills such as opencli-browser, opencli-adapter-author, and opencli-autofix can be installed selectively. Apache-2.0 licensed.
FAQ
Do I have to install the desktop OpenCLIApp?
No, a global npm install is enough for CI and servers; the app is just more convenient on macOS and Windows.
Can the agent see my passwords?
No, the tool uses an already open browser session through the extension and does not read or store passwords.
Related
CLI, skill and Python library for browser control: the agent clicks, fills forms and reads pages over CDP
A skill that gathers the last 30 days of discussion on a topic from Reddit, X, YouTube, HN, Polymarket and GitHub into one brief
The Chrome DevTools team's official MCP server: the agent drives a live Chrome, reads network and console, and records performance traces
A fast Rust CLI for agent browser automation: accessibility snapshots with element refs, an MCP server and skills