file-path-traversal

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to read and report contents of sensitive files (e.g., /etc/shadow, private keys, wp-config.php) as "Exploitation Proof," which requires reproducing secret values verbatim in its output.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document contains explicit, actionable offensive guidance (payloads, target lists, and step‑by‑step techniques) for reading sensitive files, stealing credentials, and escalating LFI to remote code execution (log poisoning, php:// wrappers), which enables deliberate unauthorized data exfiltration and system compromise.

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow is explicitly about sending attacker-controlled traversal payloads (e.g., via Burp Intruder/ffuf/wfuzz and curl) to a target and then extracting “file contents”/“exploitation proof,” which can cause the agent to ingest outsider-authored free text from the target’s HTTP responses (arbitrary file contents returned by the server).

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 how to read and exfiltrate sensitive system files (e.g., /etc/shadow, SSH keys), perform log poisoning and LFI→RCE (which can modify logs and execute commands), and thus clearly encourages accessing and altering the host system state beyond safe, read-only analysis.