hacksmarter-labs

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes examples that require embedding plaintext credentials into commands (e.g. -p '' and explicit "admin:admin"), which would force an LLM to handle or emit secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The document contains explicit, actionable instructions for credential theft, remote code execution, persistence/backdoor connections, exfiltration, and disabling defenses — techniques indicative of deliberate malicious behavior despite being framed as lab guidance.

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

• Third-party content exposure detected (medium risk: 0.30). The only “outsider-authored” free text intake is via a GitBook “agent-query endpoint” (?ask=<question>) appended by the user to a specific lab page, and the required workflow would ingest that ask payload at runtime.

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 instructs modifying system files (appending /etc/hosts with sudo), creating new local user accounts (net user ... /add), adding tun interfaces and routes with sudo, and disabling Defender—actions that change the host state and require elevated privileges.