core
The core usage guide for agent-browser, Vercel Labs' command-line tool that lets an agent drive Chrome over the DevTools protocol, and the skill the agent should read before any other agent-browser command.
What the skill teaches
The core usage guide for agent-browser, Vercel Labs' command-line tool that lets an agent drive Chrome over the DevTools protocol, and the skill the agent should read before any other agent-browser command. The basic loop is: open a page, take an interactive snapshot of its accessibility tree where each element gets a short reference, act on those references, and snapshot again after changes, preferring snapshots to screenshots. It covers navigation, clicking and filling forms, locators, waiting on specific conditions, data extraction, tabs, frames and dialogs, named sessions per worktree, credentials kept in an auth vault, WebMCP tools offered by pages, and troubleshooting. Browser output is treated as untrusted data.
Capabilities
- Open pages and act on snapshot element references
- Fill forms, click and upload files
- Wait for elements, text or URLs
- Keep separate browser sessions per task
About
- Publisher
- Vercel Labs
Install
- Run this to add it to Any client:
npm i -g agent-browser && agent-browser install
What's inside
- references/authentication.md
- references/commands.md
- references/profiling.md
- references/proxy-support.md
- references/session-management.md
- references/snapshot-refs.md
- references/streaming.md
- references/trust-boundaries.md
- references/video-recording.md
- references/webgpu.md
- templates/authenticated-session.sh
- templates/capture-workflow.sh
- templates/form-automation.sh
MCP servers and skills can run code on your computer and act on your data. Review the source before installing.
This skill includes scripts that your agent may run.
Details
| Fact | Value | Source |
|---|---|---|
| Type | Skill | |
| Tool | Browser automation | |
| Publisher | Vercel Labs | source |
| Official status | Not stated in its documentation | |
| Needs an API key or sign-in | Not stated in its documentation | |
| Includes scripts | Yes | source |
| Platforms | Not stated in its documentation | |
| AI clients | Not stated in its documentation | |
| Licence | Apache-2.0 | source |
| Pricing | Not stated in its documentation | |
| Activity | Last commit Nov 11, 202649,584 GitHub starsDownload data not available (not on npm or PyPI) | source |
| Links | Repository Website |
Copies of this skill in other repositories (3)
These differ from this skill's text. They may be adapted copies or older versions.
More from Vercel Labs
18 skills
- deploy-to-vercel (Skill)
- vercel-composition-patterns (Skill)
- vercel-react-best-practices (Skill)
- vercel-react-native-skills (Skill)
- vercel-react-view-transitions (Skill)
- web-design-guidelines (Skill)
Also listed under: Browser automation
Sources
- Skill file (SKILL.md)github.com/vercel-labs/agent-browserNov 10, 2026
- GitHub repositorygithub.com/vercel-labs/agent-browser · Backs: Licence, Publisher, ActivityNov 11, 2026
- Folder in the repository (core)github.com/vercel-labs/agent-browser · Backs: Includes scriptsNov 10, 2026