The Agent Skills Directory

[COMMAND_EXECUTION]: The scan_vulnerabilities.py script executes the git command locally to identify staged files. This is implemented using subprocess.run with a hardcoded list of arguments, which prevents command injection into the scanner itself.
[DATA_EXPOSURE]: The tool reads local source code files for analysis. It includes path validation logic that ensures user-provided directories or output paths remain within the current working directory, mitigating potential path traversal risks against the scanner's own file operations.
[SAFE]: No malicious patterns, obfuscation, or data exfiltration behaviors were detected. The skill relies exclusively on the Python standard library and operates entirely within the local environment without network dependencies.

security-scan