tailnet-multi-machine

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs reading arbitrary remote files and rewriting them via SSH heredocs (embedding file contents into generated commands/output), which can force the agent to include secrets like SSH keys, API tokens, or passwords verbatim in its output.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.95). The skill explicitly grants unilateral authority to access and modify all Tailnet machines, instructs disabling SSH host key verification, using random high ports for stealth, and creating persistent per-user systemd services — behaviors that enable backdoors, remote code execution, lateral movement, persistence, and easy data exfiltration.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly grants full operational authority across machines and instructs the agent to edit configs, install packages, manage systemd services, transfer files, create/modify system files, and to automatically accept SSH host key changes—actions that directly modify and can compromise machine state and bypass security mechanisms.