The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The framework described in the skill handles the ingestion and processing of external JSON schemas and data, which constitutes an attack surface for indirect prompt injection.\n
Ingestion points: The skill documents several entry points for external data, including createLazyView which fetches schemas from remote URLs (documented in references/schema.md), Runtime.$api which retrieves data from external services to populate stores (documented in references/stores.md), and DynamicView which renders schemas provided through component props (documented in references/components.md).\n
Boundary markers: The documentation does not mention the use of delimiters or explicit instructions to the agent to ignore potentially malicious content embedded within the fetched schemas or data.\n
Capability inventory: The skill allows for network operations via HttpClient (Runtime.$api), local data persistence via the persist configuration in stores, and dynamic UI rendering and interaction through createView and binding expressions.\n
Sanitization: There is no evidence in the provided documentation of sanitization, validation schemas, or escaping mechanisms applied to external content before it is processed or rendered by the agent.

grafton-view-dev