linux-server-audit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks for SSH passwords and shows using sshpass with the password embedded in the command (sshpass -p 'PASSWORD' ...), which requires the LLM to handle and output secret values verbatim, creating an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly instructs the agent to connect via SSH to user-supplied servers (Paso 1: Identificar Servidor) and execute/read system logs, configs and command output across the 7 audit phases (Paso 2), meaning it ingests and acts on arbitrary third-party server content that could contain malicious instructions.

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

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). Flagged because the skill explicitly instructs using sshpass and ssh with StrictHostKeyChecking=no (bypassing SSH host-key verification) and directs running system-level inspection commands over SSH, which is a security-bypassing behavior even though it does not instruct creating users or modifying files.