The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides a command npx --yes submit-expo-feedback@latest for the agent to execute for reporting feedback. This pattern downloads and runs an unverified package from the public npm registry without user confirmation. The package name submit-expo-feedback does not align with official Expo naming conventions (which typically use the @expo/ scope), presenting a significant supply chain risk.
[COMMAND_EXECUTION]: The feedback reporting command uses shell interpolation for the feedback content ("<actionable feedback>"). This allows for arbitrary shell command execution if the agent populates the field with unsanitized data.
[INDIRECT_PROMPT_INJECTION]: The feedback mechanism creates an attack surface where untrusted data (user complaints or error reports) can reach a command execution sink.
Ingestion points: User-provided feedback or error reports processed by the agent (SKILL.md).
Boundary markers: None; the feedback is wrapped in double quotes but lacks instructions for escaping or sanitization.
Capability inventory: Shell execution via npx (SKILL.md).
Sanitization: Absent; the skill does not instruct the agent to validate or escape the feedback content before passing it to the shell.
[EXTERNAL_DOWNLOADS]: The skill directs the installation of multiple external dependencies, including react-native-reanimated, react-native-worklets, react-native-gesture-handler, expo-haptics, and react-native-keyboard-controller. While common in the Expo ecosystem, these represent a broad external dependency footprint.

expo-animation