emulation-technique
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted artifacts, such as binaries, firmware images, and shellcode, for dynamic analysis and emulation.
- Ingestion points: Untrusted artifacts (ELF, PE, Mach-O, firmware images, rootfs) are processed across all skill files.
- Boundary markers: The skill includes environment-level safety gates (isolation, restricted networking, no shared secrets) in
SKILL.md. However, it does not provide explicit prompt-level delimiters or instructions to prevent the agent from inadvertently following instructions embedded in strings extracted from the targets. - Capability inventory: The skill uses powerful tools for execution and instrumentation, including QEMU, Qiling, Renode, and Frida, which are capable of running code from the untrusted artifacts.
- Sanitization: While the methodology emphasizes execution safety through isolation, it lacks specific sanitization or validation rules for the text-based data (strings, config files) extracted from analyzed artifacts before they are processed by the agent's logic.
Audit Metadata