AI agents invoke call-llm to trigger actions in Cross-LLM MCP Server. 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.
This tool executes calls to external LLM provider APIs, which constitutes triggering external operations. Misuse could result in prompt injection attacks, data exfiltration via LLM responses, unexpected API costs, or chained AI actions. The blast radius is high because the tool can route arbitrary prompts to external services, and the response could be used to drive further agent actions.
From the tool's definition 'Call a specific LLM provider by name' — triggers an external API call to a third-party LLM service
Documented attack patterns abuse exactly the kind of access call-llm gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Cross-LLM MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for call-llm:
{
"version": "1",
"default": "deny",
"tools": {
"call-llm": {
"limits": [
{
"counter": "call-llm_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}call-llm 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.
Free to start. No card required.
Call a specific LLM provider by name. It is categorised as a Execute tool in the Cross-LLM MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Cross-LLM MCP Server MCP server in PolicyLayer and add a rule for call-llm: 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 Cross-LLM MCP Server. Nothing to install.
call-llm is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the call-llm 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 call-llm. 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.
call-llm is provided by the Cross-LLM MCP Server MCP server (jamesanz/cross-llm-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Cross-LLM MCP Server, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
23 Cross-LLM MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.