The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes sudo within Tools/MassScan.ts to run the masscan utility. This is necessary for raw packet operations but enables the skill to execute commands with root privileges.\n- [EXTERNAL_DOWNLOADS]: The Workflows/UpdateTools.md workflow provides instructions for updating security tools using the pdtm utility, which downloads and installs binaries from ProjectDiscovery's official repositories.\n- [PROMPT_INJECTION]: The Workflows/AnalyzeScanResultsGemini3.md workflow is vulnerable to indirect prompt injection. It takes raw output from network scans (e.g., nmap or masscan), which may contain attacker-controlled strings such as service banners or hostnames, and includes them directly in a prompt for AI analysis.\n
Ingestion points: Scan results from files are read in Workflows/AnalyzeScanResultsGemini3.md.\n
Boundary markers: The prompt uses markdown code blocks for scan results but lacks explicit instructions for the AI to disregard any commands or directions embedded within that untrusted data.\n
Capability inventory: The skill possesses extensive network communication and shell command execution capabilities through its integrated toolset.\n
Sanitization: No filtering, escaping, or sanitization is performed on the raw scan data before it is sent to the LLM.

Recon