AI agents invoke serve to trigger actions in Ollama 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 a command to start an external service (Ollama server), which is a code/process execution action. While not destructive, it modifies system state by launching a long-running service. The severity is high because an uncontrolled server startup could consume resources, expose ports, or create security vulnerabilities if the AI agent starts it in inappropriate contexts or without proper oversight.
From the tool's definition Tool name 'serve' with description 'Start Ollama server' - initiates a server process which is an external operation with effects that depend on system state and configuration. Starting a server is an executable action that triggers system-level operations.
Documented attack patterns abuse exactly the kind of access serve gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Ollama MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for serve:
{
"version": "1",
"default": "deny",
"tools": {
"serve": {
"limits": [
{
"counter": "serve_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}serve 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.
Start Ollama server. It is categorised as a Execute tool in the Ollama MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Ollama MCP Server MCP server in PolicyLayer and add a rule for serve: 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 Ollama MCP Server. Nothing to install.
serve 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 serve 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 serve. 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.
serve is provided by the Ollama MCP Server MCP server (nighttrek/ollama-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 10 Ollama MCP Server tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
10 Ollama MCP Server tools catalogued and risk-classified — across an index of 42,500+ MCP servers.