AI agents call email_get_attachment to retrieve information from M365 without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool appears to retrieve email attachments, which is fundamentally a read operation with no inherent write or destructive capability. However, severity is elevated to medium because attachments may contain sensitive data (documents, credentials, PII), and an AI agent could exfiltrate confidential information.
From the tool's definition Tool name 'email_get_attachment' indicates retrieval of email attachment data. No description provided, but the naming convention strongly suggests a read operation that fetches attachment content from Outlook.
Attacks that exploit this kind of access
email_get_attachment. It is categorised as a Read tool in the M365 MCP Server, which means it retrieves data without modifying state.
Register the M365 MCP server in PolicyLayer and add a rule for email_get_attachment: 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 M365. Nothing to install.
email_get_attachment 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 email_get_attachment 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 email_get_attachment. 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.
email_get_attachment is provided by the M365 MCP server (robin-collins/m365-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 →