The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: Analysis of the automated scan alert confirms a false positive. The skill uses curl to fetch data from the target site which is then piped to a local Python command (python3 -c) for data processing and JSON parsing. The remote content is processed as data and is not executed by the shell or the Python interpreter.
[DATA_EXPOSURE_AND_EXFILTRATION]: The instructions include payloads for testing path traversal and local file inclusion (e.g., accessing /etc/passwd) against misconfigured development servers. These are standard security testing techniques and do not involve unauthorized access to the local agent's environment or credentials.
[DYNAMIC_EXECUTION]: The skill uses python3 -c to parse JSON and search text using regular expressions. The logic is hardcoded within the command and does not dynamically generate or execute code based on untrusted external inputs.

hunt-nextjs