binary-mobile-reversing
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill suggests executing various third-party security tools such as apktool, jadx, frida, objection, Ghidra, and binwalk to analyze binary files and mobile applications. It also describes a workflow for extracting Javascript logic and executing it via a local node environment to deobfuscate application code.\n- [EXTERNAL_DOWNLOADS]: The skill mentions identifying and using external download links for application packages, including the use of proxy services and decrypting API responses to find the original source of binary files.\n- [PROMPT_INJECTION]: The skill facilitates the ingestion of external, untrusted binary data and script source code into the agent's context for analysis. This creates an attack surface for indirect prompt injection as these files are processed using powerful system utilities and debuggers without explicit input sanitization or protective boundary markers.\n
- Ingestion points: Target files such as APKs, EXEs, and obfuscated Javascript logic (e.g., app-service.js).\n
- Boundary markers: There are no specific instructions to wrap extracted content in protective delimiters.\n
- Capability inventory: The skill utilizes command-line tools including apktool, jadx, node, frida, and pwntools.\n
- Sanitization: No sanitization or filtering of extracted strings is described before they are processed by the agent.
Audit Metadata