web-to-native
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various system and developer commands to manage the migration process, including
npmfor package installation,xcrun simctlfor iOS simulator interaction, andadbfor Android device management. - [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the
agent-browser(by Vercel Labs) andargent(by Software Mansion) utility tools. These packages are sourced from reputable organizations and are standard for the described UI verification tasks. - [PROMPT_INJECTION]: The verification workflow involves the ingestion of external data from web URLs, creating a surface for Indirect Prompt Injection.
- Ingestion points: External web content is retrieved using the
agent-browser openandsnapshotcommands inreferences/verify-on-device.md. - Boundary markers: There are no explicit boundary markers or instructions to isolate the ingested web data from the agent's control logic.
- Capability inventory: The skill context has access to command execution capabilities including
npm,xcrun, andadbacross the provided scripts. - Sanitization: No sanitization or filtering of external web content is performed before it is processed by the agent.
Audit Metadata