payload-generation

Fail

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: CRITICALCOMMAND_EXECUTIONOBFUSCATIONPERSISTENCEPRIVILEGE_ESCALATIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill contains a massive repository of shell one-liners and scripts for executing commands across multiple languages (Bash, Python, Perl, PHP, Ruby, PowerShell, Java, Lua, AWK) and operating systems.
  • [OBFUSCATION]: The skill demonstrates and uses obfuscation techniques including Base64 encoding of reverse shell code in payloads.md and the use of the Right-to-Left Override (RTLO) Unicode character (U+202E) in guides/reverse-shell-complete-guide.md to spoof file extensions.
  • [REMOTE_CODE_EXECUTION]: Instructions guide the agent to download and install third-party exploit tools from unverified GitHub repositories, such as hoaxshell, nishang, and donut. It also details 'download-and-execute' cradles using certutil, bitsadmin, and PowerShell.
  • [PERSISTENCE]: The skill provides specific commands and configurations for maintaining long-term access to a target system via Cron jobs, systemd service creation, and Windows Registry 'Run' keys.
  • [PRIVILEGE_ESCALATION]: Commands for modifying system-level settings and installing services (sudo, reg add, sc create) are provided to ensure elevated access is maintained.
  • [DATA_EXFILTRATION]: The skill's primary focus is establishing command-and-control (C2) channels, which inherently facilitates data exfiltration. It includes specific payloads for reverse TCP/HTTPS tunnels.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: User-supplied variables (LHOST, LPORT) are interpolated into executable shell commands across all guide and payload files.
  • Boundary markers: None. The instructions lack delimiters or safety warnings to prevent the agent from executing malicious variations of these inputs.
  • Capability inventory: Extensive use of Bash tool to execute generated payloads and Write tool to deploy them.
  • Sanitization: None. User inputs are directly concatenated into command strings.
Recommendations
  • CRITICAL: 5 infected file(s) detected - DO NOT USE
  • AI detected serious security threats
  • Contains 10 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 19, 2026, 04:36 PM