skills/expo/skills/expo-web-to-native/Gen Agent Trust Hub

expo-web-to-native

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs external command-line utilities for app verification and feedback submission. It fetches the agent-browser tool from Vercel's official repository and the argent simulator driver from Software Mansion. Additionally, it uses npx to execute the vendor-provided submit-expo-feedback package.
  • [COMMAND_EXECUTION]: The skill utilizes several standard development and system tools to facilitate the migration process. This includes using npm and npx for package management, as well as xcrun simctl and adb for controlling iOS and Android simulators and devices.
  • [INDIRECT_PROMPT_INJECTION]: The skill maintains an attack surface for indirect prompt injection by ingesting data from the user's repository and live web browser output. \n
  • Ingestion points: Reads the project's source code files during the initial assessment and captures accessibility trees, text, and structure from web routes using agent-browser during verification. \n
  • Boundary markers: The instructions do not define clear delimiters or specific warnings to ignore instructions that might be embedded within the processed code or web content. \n
  • Capability inventory: The skill possesses capabilities to execute shell commands, manage system processes via simctl/adb, and install global Node.js packages. \n
  • Sanitization: No validation or filtering mechanisms are applied to the data retrieved from external browser snapshots or the local file system before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 02:58 AM
Security Audit — agent-trust-hub — expo-web-to-native