【内网/内部用户】更新内部用户角色、预算、可访问模型。关键词:修改用户权限、调整预算
AI agents use litellm_user_update to create or update resources in Litellm Admin — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Litellm Admin environment.
This is a Write operation because it creates or modifies user data (roles, budgets, model access) in a reversible manner. Severity is high because misconfiguration could grant unauthorized access to models or exhaust budgets for other users, affecting multiple internal users.
From the tool's definition Tool name 'litellm_user_update' and description indicate it 'updates internal user roles, budgets, accessible models' (修改用户权限、调整预算 = modify user permissions, adjust budgets). This reversibly modifies user configuration and access controls.
Attacks that exploit this kind of access
【内网/内部用户】更新内部用户角色、预算、可访问模型。关键词:修改用户权限、调整预算. It is categorised as a Write tool in the Litellm Admin MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Litellm Admin MCP server in PolicyLayer and add a rule for litellm_user_update: 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 Litellm Admin. Nothing to install.
litellm_user_update 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 litellm_user_update 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 litellm_user_update. 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.
litellm_user_update is provided by the Litellm Admin MCP server (junwayne66/litellm-admin-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 →