The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its integration with external registry data.
Ingestion points: The skill utilizes mcp__shadcn__search_items_in_registries, mcp__shadcn__view_items_in_registries, and mcp__shadcn__get_item_examples_from_registries to bring content from external repositories into the agent context.
Boundary markers: The instructions do not include explicit delimiters or instructions for the agent to ignore or isolate instructions that may be present in the fetched registry items.
Capability inventory: The agent is granted Write, Edit, and Task tools, allowing it to modify the local filesystem and execute commands based on the content retrieved from external sources.
Sanitization: There are no explicit sanitization or content-validation steps defined for the data returned by the shadcn MCP tools before it is processed by the agent.

shadcn-components