The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests and processes untrusted data from the repository being audited.
Ingestion points: Files within the target directory specified via the --path argument are read into the agent's context.
Boundary markers: The instructions do not define specific markers to delimit untrusted file content from system instructions.
Capability inventory: The skill utilizes the Task tool to invoke sub-agents, and Bash, Grep, Read, and Write tools for local file operations.
Sanitization: No explicit sanitization or filtering of the audited file content is performed before processing.

security-auditor