AI agents invoke container_create to trigger actions in ChatGPT 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.
Starting a Docker container runs processes inside it, which constitutes execution of external operations. While creation alone might be Write, the fact that it also starts the container means code/processes are launched. This has high blast radius because a misconfigured or malicious container image could execute harmful workloads, consume resources, or be used for lateral movement.
From the tool's definition "Create and start a new Docker container" — the tool both creates and immediately starts a container, triggering execution of arbitrary containerized workloads.
Documented attack patterns abuse exactly the kind of access container_create gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and ChatGPT MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for container_create:
{
"version": "1",
"default": "deny",
"tools": {
"container_create": {
"limits": [
{
"counter": "container_create_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}container_create 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.
Create and start a new Docker container. It is categorised as a Execute tool in the ChatGPT MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the ChatGPT MCP Server MCP server in PolicyLayer and add a rule for container_create: 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 ChatGPT MCP Server. Nothing to install.
container_create 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 container_create 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 container_create. 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.
container_create is provided by the ChatGPT MCP Server MCP server (toowiredd/chatgpt-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from ChatGPT 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.
7 ChatGPT MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.