The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides functional templates for establishing reverse shells in Bash, Python, PHP, and PowerShell, enabling remote redirection of system control.
[REMOTE_CODE_EXECUTION]: Contains comprehensive exploit code for stack and heap buffer overflows, format string vulnerabilities, and ROP chain construction to achieve arbitrary code execution.
[DATA_EXFILTRATION]: Includes patterns for outbound network communication via raw TCP sockets and filesystem devices (e.g., /dev/tcp) to connect to external attacker-controlled endpoints.
[REMOTE_CODE_EXECUTION]: Demonstrates a high-risk dynamic execution vector using Python's unsafe pickle deserialization to execute arbitrary system commands.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface. Ingestion points: Processes external binary files and vulnerability reports. Boundary markers: No delimiters or warnings are present to isolate untrusted content. Capability inventory: Instructions utilize powerful tools like pwntools and msfvenom with significant system and network permissions. Sanitization: No evidence of validation or escaping for processed input data.

exploit-development