The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions to execute commands with elevated privileges using sudo, specifically for installing system dependencies and performing network monitoring with tcpdump.
[EXTERNAL_DOWNLOADS]: The skill guides the agent to install numerous third-party security tools and libraries from public registries including PyPI, apt, and Homebrew, as well as specific third-party GitHub repositories for specialized malware analysis tasks.
[PROMPT_INJECTION]: The skill processes untrusted external data from malware samples and network captures, creating an indirect prompt injection surface. Data enters the agent context via tool outputs (e.g., from strings or yara) without boundary markers or sanitization. This allows adversarial content in analyzed files to potentially influence agent behavior while the agent has access to shell and Python execution capabilities.

ctf-malware