compose-screen-architecture
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute standard build and test commands (e.g.,
./gradlew testDebugUnitTest) to verify implementation within the local project environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill analyzes project files to adapt to conventions, which constitutes a standard ingestion surface.\n - Ingestion points: Project files including
build.gradle,settings.gradle, and Android source files.\n - Boundary markers: None explicitly defined for processed project data.\n
- Capability inventory: The agent possesses file system access and the ability to execute Gradle commands.\n
- Sanitization: There is no explicit sanitization of data read from the project environment, which is typical for development-focused assistants.
Audit Metadata