The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, such as articles and documents provided by the user, which presents a surface for indirect prompt injection.
Ingestion points: The "Ingest" and "Synthesize" operations in SKILL.md involve reading and extracting facts from user-shared resources and conversation history.
Boundary markers: The instructions do not specify the use of delimiters or boundary markers to isolate external content from the agent's instructions.
Capability inventory: The skill involves reading files, writing markdown files to the knowledge/ directory, and using memory_search.
Sanitization: There are no specific instructions for sanitizing or validating the content of the external resources before they are processed and stored.

knowledge-wiki