path-traversal-lfi

Fail

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: HIGHPROMPT_INJECTIONCREDENTIALS_UNSAFEOBFUSCATIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill utilizes a specialized 'AI LOAD INSTRUCTION' block that directs the agent to function as an 'Expert attack' playbook, explicitly claiming that standard models lack capability in certain areas to encourage bypassing default model behaviors or safety filters.
  • [CREDENTIALS_UNSAFE]: The playbook defines high-value targets for sensitive data access, specifically pointing to credentials such as SSH private keys (/.ssh/id_rsa), AWS credential files (/.aws/credentials), and application environment files (.env) that store secrets.
  • [OBFUSCATION]: The skill documents numerous techniques for evading security detection, including double URL encoding (%252e), overlong UTF-8 sequences (%c0%af), and the use of 'Ghost Bits' (Unicode characters like 阮 and 阯) to bypass WAF filtering logic.
  • [COMMAND_EXECUTION]: It provides methodologies for achieving Remote Code Execution (RCE) on target systems via techniques like Log Poisoning (Apache, SSH, Mail), PHP Session File Poisoning, and exploitation of PHP wrappers.
  • [REMOTE_CODE_EXECUTION]: The skill references external exploitation tools and provides payloads for Remote File Inclusion (RFI) involving attacker-controlled servers.
  • [DATA_EXFILTRATION]: Detailed instructions are provided for reading and potentially harvesting sensitive configuration files like /etc/passwd, wp-config.php, and process environment variables.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Nov 9, 2026, 12:00 PM