format-string-exploitation
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill provides an AI load instruction to establish context as an exploitation expert. This is an instructional prompt for task performance and does not attempt to override safety filters or disregard system instructions.\n- [DATA_EXFILTRATION]: No access to sensitive files or unauthorized network communication is present in the skill documentation or code snippets.\n- [REMOTE_CODE_EXECUTION]: The Python code provided is intended for local security research using the 'pwntools' library and does not involve downloading or executing untrusted remote scripts.\n- [COMMAND_EXECUTION]: The skill documentation includes examples of vulnerable C functions and exploit methodologies for training purposes but does not execute any dangerous system commands.\n- [EXTERNAL_DOWNLOADS]: No external resources or packages are downloaded from untrusted or unknown sources.
Audit Metadata