The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a memory retrieval system using tools like hindsight_recall. This creates a surface where information stored in previous sessions (which may contain untrusted data) is ingested back into the agent's context.
Ingestion points: Data retrieved from the hosted memory bank in SKILL.md.
Boundary markers: The instructions do not specify explicit delimiters for recalled content to prevent the agent from interpreting it as instructions.
Capability inventory: The skill exposes hindsight_retain, hindsight_recall, and hindsight_reflect tools for managing memory state.
Sanitization: Content sanitization is managed by the underlying @tanstack/ai-memory and @vectorize-io/hindsight-client libraries.

tanstack-ai-memory-hindsight