re-binary-core
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration behaviors were detected. The skill instructions follow established safety protocols for the domain, including sandboxing for dynamic execution.
- [INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection via the processing of untrusted binary files. 1. Ingestion points: Binary analysis workflow (PE, ELF, Mach-O parsing). 2. Boundary markers: Separated triage and analysis phases with tool-based verification. 3. Capability inventory: Full range of reverse engineering capabilities including decompilation and debugging. 4. Sanitization: The methodology mandates the use of sandboxes (firejail/bwrap) and strict path validation for archive contents.
Audit Metadata