play-policy-insights
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches application metadata, descriptions, and Data Safety declarations from the official Google Play Store (play.google.com). This is used to cross-reference declared privacy practices with actual code behavior.
- [COMMAND_EXECUTION]: The orchestrator executes internal Python scripts to perform codebase scanning, triage audit goals, and synthesize findings into a final markdown report.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface because it ingests untrusted source code from the applications being audited and interpolates those snippets into prompts for sub-agents.
- Ingestion points: Static analysis of source files (.java, .kt, .xml, .gradle, etc.) in the target application directory and HTML content from the Google Play Store.
- Boundary markers: Code evidence is listed in the prompts for worker and critic agents, but there are minimal boundary markers or specific instructions to ignore embedded instructions within the code snippets.
- Capability inventory: The skill can execute local Python scripts, perform network requests to well-known domains, and write to a designated scratch directory.
- Sanitization: The custom template engine performs direct interpolation of code snippets without explicit sanitization or escaping of markdown control characters.
Audit Metadata