reverse-eng-radare2
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for executing the radare2 suite of command-line tools for binary analysis, disassembly, and patching. This includes a reference to starting a local HTTP command interface (r2http) for automation, which binds to the local loopback address (127.0.0.1).
- [EXTERNAL_DOWNLOADS]: The skill utilizes the r2pm package manager to download and install binary analysis plugins like r2ghidra and r2dec from external community repositories.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted binary files, creating a potential surface for indirect prompt injection if the agent interprets malicious content within the binary as instructions.
- Ingestion points: Binary samples (e.g., sample.exe) and scope configuration files (SKILL.md, cheatsheet.md).
- Boundary markers: None are defined for the tool's output, but the skill explicitly mitigates risk by advising the use of isolated virtual machines for untrusted samples.
- Capability inventory: The toolset allows for reading and writing files, as well as executing shell commands and managing external packages (SKILL.md).
- Sanitization: The skill relies on manual review by the analyst rather than automated sanitization of the analysis output.
Audit Metadata