The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents a potential vulnerability surface as it instructs the agent to process external application code and execute development commands, which is inherent to its primary purpose as a mobile engineering assistant.
Ingestion points: Mobile application source code (components, hooks, services) and configuration files (app.json, package.json).
Boundary markers: None identified in the instructional body for separating untrusted project data from agent instructions.
Capability inventory: Execution of shell commands via the Expo CLI (npx expo doctor, npx expo run:ios, etc.), file system read/write access for project organization, and network operations via standard mobile API development.
Sanitization: No explicit sanitization or validation of project-file content is specified before processing or executing commands.
[COMMAND_EXECUTION]: The skill recommends using standard Expo and React Native CLI tools. These commands are localized to the development environment and represent standard workflows for mobile app development.
Evidence: npx expo doctor, npx expo start --clear, npx expo run:ios, npx expo run:android.
[EXTERNAL_DOWNLOADS]: The skill lists standard, well-known libraries for React Native development in the project structure reference. These packages originate from official registries and are essential for the stated functionality.
Evidence: Dependencies include expo, expo-router, react-native-reanimated, and zustand.

react-native-expert