proxmox-full

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes explicit secret-like values (exported PVE_TOKEN and a plaintext root password used in -d) and shows them being placed directly into commands/headers, so an LLM reproducing these examples or generating similar commands would need to include secrets verbatim.

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 instructs the agent to perform privileged Proxmox operations (create/clone/start/stop/delete VMs and containers, modify storage, restore backups, convert templates, etc.) via the API, directly changing the hypervisor's state and thus capable of compromising the machine.