fetch_inbox
AI agents call fetch_inbox to retrieve information from MCP Agent Mail without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Based on naming convention and context (coordination layer with message threading and history), 'fetch_inbox' appears to retrieve or query existing messages without modifying data. This is a Read operation. Severity is low because fetching inbox data has minimal blast radius—it does not modify state, execute code, or affect financial systems.
From the tool's definition Tool name 'fetch_inbox' suggests retrieving messages/data from an inbox. The server description indicates it provides 'searchable history' and 'message threading', supporting a read operation. The empty description limits confidence.
Attacks that exploit this kind of access
fetch_inbox. It is categorised as a Read tool in the MCP Agent Mail MCP Server, which means it retrieves data without modifying state.
Register the MCP Agent Mail MCP server in PolicyLayer and add a rule for fetch_inbox: 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 MCP Agent Mail. Nothing to install.
fetch_inbox 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 fetch_inbox 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 fetch_inbox. 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.
fetch_inbox is provided by the MCP Agent Mail MCP server (omelchmichael/mcp_agent_mail). 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 →