The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install official vendor-maintained libraries and tools from standard package registries:
Node.js: @tryterra/cli and terra-react via NPM.
Android: co.tryterra:terra-android via Maven Central.
Flutter: terra_flutter_bridge via Pub.dev.
iOS: TerraiOS via Swift Package Manager (GitHub).
[COMMAND_EXECUTION]: The documentation provides examples of using the terra CLI for administrative tasks, such as managing environment data tokens and listing users. These commands are intended for development and integration management. The skill explicitly warns against printing sensitive credentials in logs when using commands that output auth tokens.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for ingesting untrusted data from third-party health sources (Apple Health, Samsung Health, Health Connect) into the agent's context via SDK getter methods like getDaily and getSleep.
Ingestion points: Data is retrieved via TerraManager (Android/iOS), TerraFlutter (Flutter), or terra-react (React Native) functions.
Boundary markers: No specific delimiters or warnings for the agent are defined within the skill's instructions for handling the returned health data.
Capability inventory: The skill includes the ability to write data to mobile health stores using postActivity, postBody, and postNutrition methods, as well as managing the integration via the terra CLI.
Sanitization: The documentation does not specify sanitization procedures for the health data, though such data (numerical metrics and enums) typically poses a low risk for prompt injection compared to free-text inputs.

terra-mobile-sdk