The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions and examples establish a pattern for ingesting external data and incorporating it into HTTP responses, which creates a surface for indirect prompt injection.
Ingestion points: The examples in references/response-basics.md and references/response-containers.md demonstrate route handlers taking arguments such as name: str and request: Request, which represent untrusted external data.
Boundary markers: The provided code snippets do not include explicit boundary markers or instructions to the agent to treat interpolated data as untrusted content.
Capability inventory: The skill utilizes capabilities including writing to HTTP response bodies, setting headers/cookies, and serving files via the File container as seen in references/response-containers.md.
Sanitization: No sanitization or validation logic is present in the examples to filter or escape external content before it is returned in the response context.

litestar-responses