AI agents use create_agent to create or update resources in Retellai — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Retellai environment.
Creating a new agent is a write operation that adds reversible state to the system. It has a medium severity because a malicious agent configuration could potentially be used to make unwanted phone calls or interact with users deceptively, but the core action itself is reversible via the delete_agent capability.
From the tool's definition Tool name 'create_agent' and description 'Creates a new Retell agent' indicate a create operation that generates new data/configuration entities.
Attacks that exploit this kind of access
Creates a new Retell agent. It is categorised as a Write tool in the Retellai MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Retellai MCP server in PolicyLayer and add a rule for create_agent: 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 Retellai. Nothing to install.
create_agent 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_agent 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_agent. 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_agent is provided by the Retellai MCP server (@abhaybabbar/retellai-mcp-server). 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 →