fastmcp-stdio-proxy

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). Contains a high-risk instruction to mount the host Docker socket (/var/run/docker.sock) and run the Docker CLI from inside a container, which effectively grants full host control (remote code execution, privilege escalation, ability to spawn containers that can exfiltrate data or persist), a pattern frequently abused as a backdoor.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow uses stdio_client to invoke docker run -i --rm <target-image> (Docker-in-Docker via the host docker.sock), which will cause the agent to read the stdout/stderr text produced by an outsider-provided container image/command as LLM-readable prose through the stdio transport.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The prompt explicitly instructs mounting the host Docker socket (/var/run/docker.sock) into a container and installing the Docker CLI so the container can spawn and control host containers, which effectively grants root-level control of the host and compromises machine state.