The Agent Skills Directory

[SAFE]: The skill provides comprehensive and standard instructions for binary analysis using well-known open-source tools (rizin and radare2). All identified patterns are consistent with the skill's stated purpose of providing a workflow for command-line analysis.
[EXTERNAL_DOWNLOADS]: The skill mentions installing rizin/radare2 via official package managers (brew, pacman, dnf, apt, choco) and using the tool's native package manager (rz-pm) to install the rz-ghidra plugin. These are standard practices for these tools and target official or well-known distribution channels.
[COMMAND_EXECUTION]: The skill documents the use of various command-line tools for analysis (rizin, rz-bin, r2, rabin2). These are standard tools for the reverse engineering domain.
[DYNAMIC_EXECUTION]: The skill describes the use of rizin's built-in debugger (rizin -d) and the installation of plugins. These features are essential for the primary use case of the skill and are documented alongside warnings regarding environment limitations (e.g., Windows debugging limitations, Linux ptrace constraints).
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process external binary samples which could theoretically be malicious, it explicitly advises performing all operations within a sandbox environment and follows standard analysis protocols, mitigating risk to the agent's host environment.

re-radare2