The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides detailed instructions for executing a wide range of offensive security workflows, including network enumeration, service discovery, and automated brute-force attacks against common protocols like SSH, FTP, and SMB.
[REMOTE_CODE_EXECUTION]: The instructions include specific command patterns for deploying known exploits (such as EternalBlue and Shellshock) and gaining interactive shells on remote systems through automated exploitation frameworks and database injection tools.
[DYNAMIC_EXECUTION]: The skill documentation describes methods for generating custom malicious payloads and reverse shell scripts at runtime for Windows, Linux, and various web environments (PHP, ASP, JSP).
[DATA_EXFILTRATION]: The skill includes command references for capturing live network traffic and extracting information from packet captures, which can be used to harvest sensitive data from network communications.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to handle untrusted identifiers such as IP addresses and URLs provided during assessment tasks.
Ingestion points: Target URLs, IP addresses, and hostname identifiers provided by the user or discovered during scans (SKILL.md).
Boundary markers: None; the instructions do not provide delimiters or instructions to prevent the agent from interpreting content found within target systems.
Capability inventory: The agent has shell access to powerful tools capable of automated exploitation and packet inspection.
Sanitization: No sanitization or validation logic is provided to ensure that target inputs are safe before being used in shell command strings.

pentest-commands