format-string-exploitation
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEPROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains an
AI LOAD INSTRUCTIONblock designed to steer the agent's behavior towards an "expert attack" context. This is intended to influence the model's logic regarding memory address handling and parameter offsets to improve exploit reliability. - [DYNAMIC_EXECUTION]: The documentation includes Python code snippets that use the
pwntoolslibrary to generate binary exploitation payloads. While provided as templates for security research, they involve the generation of executable memory-corruption code.
Audit Metadata