The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by processing and retrieving user-provided documents.\n
Ingestion points: Content is ingested through the index document endpoint as described in the API usage examples.\n
Boundary markers: The skill guidelines do not include specific instructions or delimiters to help the agent distinguish between retrieved search results and authoritative system instructions.\n
Capability inventory: The skill utilizes curl for making network requests and writes temporary files to /tmp/.\n
Sanitization: There are no instructions for sanitizing, validating, or escaping external content before it is indexed or after it is retrieved.

moss