The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (UXKIN's database of screens and design systems), which presents a theoretical surface for indirect prompt injection if the source material contains malicious instructions.
Ingestion points: Data enters the context via the find_ui_references, get_journey, and find_ui_materials MCP tools.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the data returned by tools.
Capability inventory: The skill itself has no capabilities for subprocess execution, file-writing, or direct network operations; it only processes and reports tool output.
Sanitization: No explicit sanitization or validation of the fetched design data is mentioned in the instructions.

uxkin-research