re-unpack-simple
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references official repositories and standard package managers for acquiring security tools. It directs users to well-known package managers (apt, dnf, pacman, brew, choco) and official GitHub repositories for 'upx' and 'Scylla'.
- [INDIRECT_PROMPT_INJECTION]: The skill involves processing binary files (e.g., sample.exe), which is an inherent attack surface for reverse engineering tasks. Ingestion points: External binary samples are loaded into tools like upx and debuggers like gdb or x64dbg. Boundary markers: The instructions explicitly mandate that unpacking and verification must occur within a sandbox environment. Capability inventory: The skill leverages tools like gdb, wine, and Scylla to manipulate binary code. Sanitization: No sanitization of the binary content is performed as the skill's purpose is analysis, which is appropriately mitigated by the required use of a sandbox.
- [SAFE]: No malicious patterns, such as prompt injection, hidden obfuscation, or unauthorized data exfiltration, were identified in the instructions.
Audit Metadata