AI agents use terminate_agent to create or update resources in ZMCPTools — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your ZMCPTools environment.
An AI agent can call terminate_agent faster than any human can review — one bad instruction and it creates or modifies resources in ZMCPTools by the hundred, each call as confident as the last.
Documented attack patterns abuse exactly the kind of access terminate_agent gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and ZMCPTools, and nothing reaches the server without passing your rules. This is the rule we recommend for terminate_agent:
{
"version": "1",
"default": "deny",
"tools": {
"terminate_agent": {
"limits": [
{
"counter": "terminate_agent_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}terminate_agent stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Terminate one or more agents. It is categorised as a Write tool in the ZMCPTools MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the ZMCPTools MCP server in PolicyLayer and add a rule for terminate_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 ZMCPTools. Nothing to install.
terminate_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 terminate_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 terminate_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.
terminate_agent is provided by the ZMCPTools MCP server (zachhandley/zmcptools). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 70 ZMCPTools tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
70 ZMCPTools tools catalogued and risk-classified — across an index of 42,500+ MCP servers.