| name | web-browse |
| description | Navigate to a URL, extract page content, and take screenshots |
| tags | ["browser","web","navigate","extract","screenshot"] |
| version | 1.0.0 |
Web Browse
Navigate to a URL and extract its content for the requesting agent.
When to use
- "Go to this URL and tell me what's there"
- "Get the content from this page"
- "Take a screenshot of this site"
- "What does this webpage say?"
How to use
1. Start browser (if not running)
browser start
2. Navigate
browser navigate <url>
3. Extract content
browser snapshot # Get page content as structured text
browser screenshot # Capture visual state as image
4. Stop browser when done
browser stop
Status Reporting
- On start: "Navigating to [URL] for [purpose]"
- On complete: Return extracted content + screenshot path if taken
- On failure: What broke, why, is it retryable?
Failure Handling
- If page doesn't load: report error, check if URL is valid
- If blocked by bot detection: report as known issue, escalate
- If content is behind auth: report, ask Captain if credentials are available
- Always chart new failure patterns via memory_store
Intent: Resourceful [I07]. Purpose: [P-TBD].