react-native-device-data-module

Fail

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: HIGHDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: Facilitates access to highly sensitive Personally Identifiable Information (PII) and authentication credentials.
  • The skill provides a unified mapping to access userSSN (Social Security Number), which is highly sensitive PII. This is mapped to encryptQRSSNBackNumber on iOS and USER_SSN_{userId} on Android.
  • The module exposes accessToken and refreshToken keys, which grant access to user sessions and sensitive resources.
  • Device identifiers such as deviceId (mapped to deviceToken or PREF_ANDROID_DEVICE_UUID) are enabled for retrieval, which can be used for persistent user tracking.
  • [PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface.
  • Ingestion points: The skill retrieves data from native device storage (UserDefaults/SharedPreferences), which can contain user-controlled or externally-influenced data.
  • Boundary markers: The documentation does not specify or enforce the use of delimiters or "ignore embedded instructions" warnings when processing these retrieved values.
  • Capability inventory: The skill provides the agent with a direct interface to read and transform sensitive stored values for use within the agent's logic flow.
  • Sanitization: There is no mention of input validation, escaping, or sanitization for values extracted from native storage before they are interpolated into the agent's context, potentially allowing stored data to influence agent behavior.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 9, 2026, 06:47 AM