The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous high-risk commands for sensitive system operations, such as dumping LSASS memory via procdump or comsvcs.dll and manipulating Windows services to gain SYSTEM privileges.
[DATA_EXFILTRATION]: Explicit instructions and templates are included for exfiltrating sensitive data to external attacker-controlled infrastructure using curl POST requests and SMB shares.
[REMOTE_CODE_EXECUTION]: The skill facilitates the download and execution of third-party binary exploitation tools such as Mimikatz, Rubeus, and SharpHound from unverified GitHub repositories.
[EXTERNAL_DOWNLOADS]: Multiple prerequisites involve downloading code and binaries from non-trusted external sources like lgandx/Responder and various individual GitHub accounts.
[CREDENTIALS_UNSAFE]: Includes commands for extracting high-value credentials including the SAM database, SYSTEM registry hive, and the Active Directory NTDS.dit database.

network-pentest