AI agents call list_accounts to retrieve information from Email without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only retrieves metadata about configured email accounts and their connection status. It has no side effects, does not execute operations, and does not modify data. It is a simple informational query, placing it firmly in the Read category. Severity is low because the information exposed (account names and connection status) is already known to the agent and poses minimal risk even if misused.
From the tool's definition Tool name is 'list_accounts' and description states it retrieves 'the mailbox this agent key is bound to (and whether it is connected)' — a passive query of account connection status with no modification or execution.
Attacks that exploit this kind of access
The mailbox this agent key is bound to (and whether it is connected). It is categorised as a Read tool in the Email MCP Server, which means it retrieves data without modifying state.
Register the Email MCP server in PolicyLayer and add a rule for list_accounts: 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 Email. Nothing to install.
list_accounts 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 list_accounts 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 list_accounts. 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.
list_accounts is provided by the Email MCP server (swapnilsurdi/email_mcp). 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 →