cmdi-command-injection
Fail
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONOBFUSCATION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides numerous functional reverse shell payloads for various environments including Bash, Python, Perl, Netcat, and PowerShell. These payloads are designed to grant an attacker interactive control over the host system.
- [DATA_EXFILTRATION]: Detailed instructions and templates for Out-of-Band (OOB) exfiltration are included. These demonstrate how to harvest sensitive data like user identifiers, environment variables, and system information via DNS lookups and HTTP requests to attacker-controlled infrastructure.
- [COMMAND_EXECUTION]: The document contains a complete reference of shell metacharacters and vulnerable code patterns across multiple programming languages (PHP, Python, Node.js, Perl, ASP). It is explicitly designed to teach an agent how to inject and execute arbitrary system commands.
- [OBFUSCATION]: A dedicated section outlines techniques for bypassing security filters and Web Application Firewalls (WAFs). It covers the use of character alternatives (e.g., using
$IFSor globbing to replace spaces), variable assembly, and encoding (hex, octal) to conceal malicious command strings from detection. - [PRIVILEGE_ESCALATION]: The skill documents advanced techniques to bypass restricted environments, such as escaping PHP's
disable_functionsviaLD_PRELOADand exploiting unauthenticated container runtime APIs (Docker/Kubernetes) to gain elevated access to host resources.
Recommendations
- CRITICAL: 2 file(s) identified as malware by FileRep - DO NOT USE
- AI detected serious security threats
Audit Metadata