Agent Skills
Browserbase
Cloud browser infrastructure for AI agents and automation
Browserbase is a cloud browser infrastructure service that provides managed browser sessions for AI agents and automation workflows with Playwright and Puppeteer support.
Required Environment
| Name | Type | Description |
|---|---|---|
BROWSERBASE_API_KEY | secret | API key from Browserbase Dashboard |
BROWSERBASE_PROJECT_ID | secret | Project ID from Browserbase Dashboard |
Configuration
version: "1.0"
agents:
my-agent:
framework: claude-code
skills:
- https://github.com/vm0-ai/vm0-skills/tree/main/browserbaseRun
Store your secrets on the platform (recommended, one-time setup):
vm0 secret set BROWSERBASE_API_KEY your-browserbase-api-key
vm0 secret set BROWSERBASE_PROJECT_ID your-browserbase-project-idThen run your agent - secrets are automatically loaded:
vm0 run my-agent "automate browser tasks"For CI/CD or temporary overrides, pass secrets at runtime: --secrets BROWSERBASE_API_KEY=value. See Environment Variables for details.
Example Instructions
# Browser Automation Agent
You use Browserbase to automate web browser tasks with persistent sessions.
## Workflow
1. Create a new browser session
2. Use the session with Playwright or Puppeteer
3. Perform automation tasks
4. Retrieve session logs and recordings
## Capabilities
- Create managed browser sessions
- Persist authentication across sessions using contexts
- Access live debugging URLs
- Download session recordings and logs
- Upload files for use in sessions# Web Testing Agent
You use Browserbase to run automated tests across multiple browsers.
## Workflow
1. Create browser session with specific configuration
2. Run test suite via Playwright
3. Capture screenshots and recordings
4. Generate test reports from session data
## Guidelines
- Configure session timeouts and regions as needed
- Use contexts to persist authentication state
- Enable proxy support on paid plans for protected sites
- Monitor session status and usage limits