The Agent Skills Directory

[COMMAND_EXECUTION]: The provided examples include the use of sudo masscan. Executing commands with sudo grants administrative privileges, which is necessary for certain network operations like raw packet crafting but represents a privilege escalation vector if misused.
[EXTERNAL_DOWNLOADS]: The skill references the public GitHub repository for the nerva fingerprinting tool hosted by Praetorian Inc.
[PROMPT_INJECTION]:
Ingestion points: The skill processes external data from files via the -l argument and from standard input via command pipes.
Boundary markers: There are no explicit markers used to distinguish untrusted input from the processing logic.
Capability inventory: The skill utilizes command-line tools (nerva, naabu, masscan) to perform network scanning and fingerprinting operations.
Sanitization: Input data is passed to tool arguments without documented validation or sanitization, potentially allowing for indirect prompt injection if the processed data contains malicious control sequences.

nerva-scan