android-ui-states-validation
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to review and process Android source code and layout files (e.g., in the
examples/directory), which represents a standard surface for indirect prompt injection if the analyzed code contains malicious instructions. - Ingestion points: Review of screen destinations, state holders, and UI code in
SKILL.md. - Boundary markers: None specified in the instructions to distinguish between agent commands and data being reviewed.
- Capability inventory: The skill can execute local build commands (
./gradlew) and Python scripts (scripts/eval_triggers.py). - Sanitization: No explicit sanitization or validation of the ingested code files is documented.
Audit Metadata