AI agents use create_chat to create or update resources in Teams — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Teams environment.
This tool creates (writes) new chat data in Teams, which is a reversible action. It does not execute code, delete data, or handle finances. The severity is medium because creating chats on behalf of a user could be used for spam, social engineering, or unwanted communication campaigns, but the blast radius is limited to the chat system without affecting other data or systems.
From the tool's definition Tool name is 'create_chat' and description states 'Create a new chat conversation' — explicitly creates a new communication channel with one or more users.
Attacks that exploit this kind of access
Create a new chat conversation. Can be a 1:1 chat (with one other user) or a group chat (with multiple users). Group chats can optionally have a topic. It is categorised as a Write tool in the Teams MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Teams MCP server in PolicyLayer and add a rule for create_chat: 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 Teams. Nothing to install.
create_chat is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the create_chat 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 create_chat. 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.
create_chat is provided by the Teams MCP server (@floriscornel/teams-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.