The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests external data from the Library of Congress JSON API, creating a surface for indirect prompt injection.
Ingestion points: The search and get_item methods in scripts/loc_api.py process external data retrieved from loc.gov.
Boundary markers: Data is processed as structured JSON, and the agent is provided with clear schemas for interpretation, though the content itself is not wrapped in specific security delimiters.
Capability inventory: The skill performs network GET requests to the official LOC API and includes instructions for using the gh CLI for issue reporting to the author's repository.
Sanitization: The _path_segment helper function validates input used in URL construction to prevent directory traversal attacks during item lookups.
[SAFE]: The skill follows security best practices for external API integration. It uses the standard Python library for networking, implements robust error handling with exponential backoff for retries, and strictly limits its operations to the intended primary purpose of library catalog search.

loc-catalog