The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it is designed to ingest and interpret data from untrusted external project files.
Ingestion points: The skill uses Read, Glob, and Grep tools, along with the scripts/security_scan.py script, to read and process content from a user-specified project directory.
Boundary markers: There are no explicit instructions or delimiters used to ensure the agent ignores or sanitizes potential instructions embedded within code comments, documentation, or configuration files in the scanned project.
Capability inventory: The skill has the capability to execute shell commands via the Bash tool and the subprocess module in the included Python script.
Sanitization: No sanitization or validation of the retrieved content is performed before it is presented to the agent for analysis.
[COMMAND_EXECUTION]: The scripts/security_scan.py script utilizes the subprocess.run method to execute system commands, specifically npm audit, to perform dependency vulnerability assessments.
[EXTERNAL_DOWNLOADS]: The skill performs network operations by executing npm audit, which connects to the official npm registry to fetch current vulnerability data. This is documented as a standard and safe operation for dependency auditing.

vulnerability-scanner