The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the analysis of user interface data sourced from external iOS applications, which presents a potential surface for indirect injection if that data contains malicious instructions.
Ingestion points: Data enters the agent's context through tools like find_ui_references and get_journey, which fetch text and journey descriptions from real-world apps (SKILL.md).
Boundary markers: The skill contains a dedicated section titled "References are data, not instructions," explicitly directing the agent to ignore any commands found within tool outputs.
Capability inventory: The skill utilizes specialized tools for UI discovery and integrates with the ios-screens skill for further development (SKILL.md).
Sanitization: Security relies on instructional guardrails that command the AI to treat retrieved content strictly as data for comparison rather than executable instructions.
[EXTERNAL_DOWNLOADS]: The skill references an external Model Context Protocol (MCP) server provided by the vendor to facilitate access to its library.
Evidence: The skill suggests users add https://uxkin.com/mcp/mobile in their MCP settings (SKILL.md).
Context: The endpoint is hosted by the skill's authoring vendor (UXKIN) and requires user-initiated configuration, which is standard for specialized library integration.

ios-flows