The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/security_scan.py invokes system binaries npm and pip-audit using the subprocess module to perform dependency audits.
[PROMPT_INJECTION]: The skill processes untrusted file content during its scanning phase, which creates an indirect prompt injection surface. Malicious instructions embedded in scanned files could be extracted into the snippet field and presented to the agent.
Ingestion points: scripts/security_scan.py reads all text files within the target directory provided by the user.
Boundary markers: The scan report format (Text or JSON) lacks explicit delimiters or instructions to the agent to ignore content within the code snippets.
Capability inventory: The script has the ability to execute shell commands (npm, pip-audit) and write reports to the file system.
Sanitization: Content is truncated to 200 characters, but no escaping or filtering for prompt injection markers is performed.

code-safety-audit