The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external artifacts like code repositories and documents, which creates a potential surface for indirect injection.
Ingestion points: File repositories, diffs, and documents.
Boundary markers: Explicit instructions to treat artifacts as evidence only.
Capability inventory: Shell execution for searching and file writing.
Sanitization: Instructions to keep real values out of reports and use data categories instead.
[DYNAMIC_EXECUTION]: The skill generates shell command templates using ripgrep for local data discovery. This is standard and appropriate tool usage for identifying data fields during a privacy audit.
[EXTERNAL_DOWNLOADS]: The skill references authoritative legislative trackers and technical specifications from trusted organizations, including the IAPP, W3C, and state Attorney General offices (Colorado, New Jersey). These references are used for informational purposes and do not involve remote code execution.

us-state-privacy-compliance