The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands for Android development and process management.
Uses npx react-native run-android and npx expo run:android to build and launch applications.
Employs adb commands (pm list packages, dumpsys window, logcat, shell input keyevent, reverse tcp, am force-stop, am start) for device interaction and log capturing.
Uses lsof and kill to manage stale Metro bundler processes on port 8081.
These operations are consistent with the skill's primary purpose of mobile application testing.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from external files and runtime logs.
Ingestion points: Reads package.json, app.json, app.config.js, android/app/build.gradle, and adb logcat logs to extract configuration and detect errors.
Boundary markers: The instructions do not specify boundary markers or delimiters when parsing these external sources.
Capability inventory: The skill can execute shell commands via adb and npx, and it uses bundled scripts (launch.sh, snap.sh, stop.sh) to automate actions.
Sanitization: No explicit sanitization of the content read from configuration files or logs is mentioned.
While the surface exists, it is characteristic of development and testing tools and does not represent an active threat in this context.

react-native-test-android-visually