The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts scripts/vapt-recon.py and scripts/vapt-chain.py utilize subprocess.run(..., shell=True) to orchestrate various security tools. User-supplied inputs, including targets, usernames, and passwords, are interpolated directly into shell command strings without sanitization. This allows for arbitrary command injection if an attacker provides a malformed target string (e.g., ; malicious_command).
[REMOTE_CODE_EXECUTION]: The reference file references/exploitation-chaining.md contains multiple functional reverse shell payloads in Python and Bash. These payloads are designed to establish outbound network connections and provide interactive shell access to a remote host. These patterns were correctly identified by YARA signature scanning as high-risk execution indicators.
[EXTERNAL_DOWNLOADS]: The installation guide in references/metasploit-reference.md includes a workflow that downloads a script from a remote URL (raw.githubusercontent.com/rapid7/...) and executes it. While the source belongs to a well-known security organization (Rapid7), the practice of piping remote content directly to a shell for installation constitutes a remote code execution vector.

vapt