styles

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill uses established Android development practices and references official documentation from Google's Android developer portals and GitHub repositories. No malicious patterns were identified.
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with the local development environment by processing project files and instructing the agent to execute build and test commands.
  • Ingestion points: The agent reads local Android project source code, including Gradle configurations (build.gradle.kts) and UI component files, to identify migration targets.
  • Boundary markers: No specific delimiters are used to separate the project's source code from the instructions provided to the agent.
  • Capability inventory: The skill involves modifying existing code files and executing shell-based development tools for running Android emulators and UI tests (Espresso/UI Automator).
  • Sanitization: As a development aid, the skill operates on the user's own codebase and does not implement external data sanitization.
  • [COMMAND_EXECUTION]: The migration workflow requires running visual regression tests and Android emulators to verify component styling.
  • Evidence: Instructions in Step 3 guide the execution of screenshot tests using standard Android testing frameworks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 01:15 PM
Security Audit — agent-trust-hub — styles