The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a potential attack surface for indirect prompt injection.
Ingestion points: The skill uses WebFetch, mcp__scout__navigate, and mcp__claude_ai_Figma__get_design_context to ingest content from live websites and Figma designs as specified in the allowed-tools and instructions.
Boundary markers: The instructions do not specify any delimiters or warnings to the agent regarding the treatment of embedded instructions within the analyzed designs or web pages.
Capability inventory: The skill utilizes Read, Glob, Grep, and web-browsing tools. While it lacks destructive capabilities like file writing or arbitrary code execution, its output is used to influence developer and designer decisions.
Sanitization: The instructions do not define any sanitization, escaping, or filtering logic for the external content before it is interpolated into the agent's context for analysis.

ux-expert