completion · Litellm MCP: Risk & Policy

WHAT IT DOES

What completion does on Litellm

AI agents invoke completion to trigger actions in Litellm. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.

WHY IT NEEDS A POLICY

Why completion needs a policy

This tool invokes an external AI model inference API. While it doesn't destructively modify data, it executes external operations (LLM inference) that incur API costs, consume resources, and produce outputs whose nature depends entirely on the prompt arguments. It fits Execute rather than Read because it triggers a stateful external operation rather than merely querying local data.

From the tool's definition "Generate text completions" using LiteLLM — triggers external LLM API calls whose effects (costs, content generated) depend on the arguments supplied.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

completion Rate-limited

completion stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.

View as policy code

policy.json

{
  "version": "1",
  "default": "deny",
  "tools": {
    "completion": {
      "limits": [
        {
          "counter": "completion_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

RATE-LIMIT THIS TOOL →

Instant setup, no code required.

PolicyLayer is an MCP gateway: every tool call is checked against your rule before it reaches Litellm, so you can run agents against it safely.

FAQ

Questions about completion

What does the completion tool do? +

Generate text completions (legacy endpoint) using LiteLLM. It is categorised as a Execute tool in the Litellm MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on completion? +

Register the Litellm MCP server in PolicyLayer and add a rule for completion: 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. Nothing to install.

What risk level is completion? +

completion is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit completion? +

Yes. Add a rate_limit block to the completion 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.

How do I block completion completely? +

Set action: deny in the PolicyLayer policy for completion. 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.

What MCP server provides completion? +

completion is provided by the Litellm MCP server (litellm-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.