AI agents call browser_read to retrieve information from Byob without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves webpage content for the AI to consume. It performs a passive query operation analogous to fetching a web page—no data is created, modified, deleted, or executed. The action is purely informational with no side effects beyond reading. Low severity because the blast radius of misuse is minimal: reading accessible web pages poses no direct risk to systems or data integrity.
From the tool's definition Tool name 'browser_read' and description 'Read full content from a webpage' indicate data retrieval without modification or side effects.
Attacks that exploit this kind of access
Read full content from a webpage using the user. It is categorised as a Read tool in the Byob MCP Server, which means it retrieves data without modifying state.
Register the Byob MCP server in PolicyLayer and add a rule for browser_read: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Byob. Nothing to install.
browser_read is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the browser_read rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for browser_read. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
browser_read is provided by the Byob MCP server (wxtsky/byob). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.