AI agents call read_notifications to retrieve information from Social without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves existing notification data from Facebook without creating, modifying, deleting, or executing any operations. It has no side effects beyond reading information the authenticated user already has access to. Blast radius is minimal—misuse could only expose the user's own notifications to the AI agent.
From the tool's definition Tool name 'read_notifications' and description 'Fetch your Facebook notifications' explicitly indicate data retrieval with no modification. The action is a query operation that returns notification data as a markdown table.
Attacks that exploit this kind of access
Fetch your Facebook notifications as a markdown table. It is categorised as a Read tool in the Social MCP Server, which means it retrieves data without modifying state.
Register the Social MCP server in PolicyLayer and add a rule for read_notifications: 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 Social. Nothing to install.
read_notifications 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 read_notifications 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 read_notifications. 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.
read_notifications is provided by the Social MCP server (whypuss/ai-cdp-browser). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →