detox
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates mobile automation workflows that can ingest untrusted data from external sources.
- Ingestion points: Data entering through application URI schemes via the
device.openURLmethod and UI element properties inSKILL.mdserves as a potential vector for external content. - Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to treat data retrieved from the application or deep links as untrusted.
- Capability inventory: The skill provides broad capabilities for mobile device interaction, including application lifecycle management, UI navigation, and automated typing.
- Sanitization: No sanitization or verification mechanisms are described for inputs processed during the testing or deep linking flows.
Audit Metadata