The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides detailed command-line examples for using offensive security tools such as NetExec (nxc), Kerbrute, and Hydra. These instructions are intended to guide an agent in performing authorized authentication attacks against services like SMB, LDAP, and Kerberos in a penetration testing context.- [INDIRECT_PROMPT_INJECTION]: The skill describes processes that involve ingesting and acting upon external, potentially untrusted data, such as harvested user lists and breach credentials. While this represents a theoretical vulnerability surface for indirect prompt injection, it is an inherent part of the penetration testing task the skill is designed to perform.
Ingestion points: External user lists (e.g., users.txt) and harvested credentials referenced in the instructions.
Boundary markers: The instructions do not define specific delimiters or isolation techniques to prevent the agent from interpreting data within the user lists as instructions.
Capability inventory: The skill outlines capabilities for executing shell commands and performing network-based authentication attempts using tools like nxc and kerbrute.
Sanitization: There are no instructions for validating or sanitizing the input data from the user lists before processing.

network-password-spraying