The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill derives testing logic from project files which could contain malicious instructions. * Ingestion points: The agent reads PRODUCT.md to define its core-loop test journey and monitors application logs for health checks. * Boundary markers: Instructions limit the agent to reading the file only and establish explicit logical gates, though technical delimiters are not used. * Capability inventory: The skill utilizes shell access and the agent-device CLI for device interaction and system state manipulation. * Sanitization: The skill does not perform validation or sanitization of the content derived from PRODUCT.md before processing it.
[COMMAND_EXECUTION]: The skill relies on shell commands for building native binaries, managing mobile emulators, and executing automated test scripts.
[EXTERNAL_DOWNLOADS]: The skill prompts the installation of the agent-device utility via npm to satisfy functional requirements for app verification.

expo-smoke-test