re-tee
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides legitimate technical instructions for reverse engineering TEE/TrustZone components, including OP-TEE and Trusted Applications. It follows security best practices by recommending that all dynamic analysis of host-side components be performed within a sandbox environment.
- [INDIRECT_PROMPT_INJECTION]: The skill involves processing untrusted external data such as firmware images, TEE OS binaries, and host-side logs, which is the primary purpose of a reverse engineering tool.
- Ingestion points: Firmware extraction from bootrom or storage partitions (SKILL.md), binary parsing of .ta (Trusted Application) files, and observation of host-side ioctl parameters and system logs.
- Boundary markers: The instructions focus on technical extraction and documentation but do not explicitly define delimiters for raw data parsed from analyzed binaries.
- Capability inventory: The skill utilizes powerful external analysis tools including Ghidra, IDA Pro, and Frida for code inspection and system monitoring.
- Sanitization: No specific prompt-level sanitization for ingested binary strings is mentioned, as the data is intended for expert technical analysis.
Audit Metadata