The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous command-line templates for security auditing tools, including sqlmap, curl, and jwt_tool.py. These include high-risk flags such as --os-shell in sqlmap, which is designed to provide shell access on a target system.
[DATA_EXFILTRATION]: Provides instructional templates for exfiltrating data via out-of-band channels, such as DNS lookups (nslookup) and HTTP POST requests (curl) to external domains. These are intended for verifying blind command injection vulnerabilities.
[PROMPT_INJECTION]: The skill possesses a significant attack surface for indirect prompt injection due to its core functionality.
Ingestion points: The skill is designed to process untrusted external data, specifically target URLs and HTTP request files (referenced as request.txt).
Boundary markers: No delimiters or instructions are provided to the agent to distinguish between target data and its own security logic.
Capability inventory: The skill utilizes powerful system capabilities, including full network access via curl and sqlmap, and the execution of external Python scripts.
Sanitization: There is no evidence of sanitization, validation, or filtering of the responses received from external web targets before they are processed by the agent.

web-pentest