re-ctf
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a legitimate technical framework for Reverse Engineering and CTF tasks, utilizing standard tools like angr, z3, and Ghidra for binary analysis.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection given that it processes untrusted binary challenge files.\n
- Ingestion points: External CTF binaries and challenge scripts processed in SKILL.md.\n
- Boundary markers: The instructions explicitly mandate the use of [[re-sandbox]] for verification and solver execution.\n
- Capability inventory: Binary analysis and execution via gdb, radare2, and symbolic execution via angr/z3.\n
- Sanitization: Employs sandboxing and initial triage ([[re-triage]]) to isolate untrusted data.\n- [SAFE]: The guidance for analyzing PyInstaller binaries, including frozen module comparison and marshal object hashing, represents established professional forensic techniques.
Audit Metadata