google-adk
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of instructional content and code examples for the Google Agent Development Kit (ADK). It does not contain executable malicious code, obfuscation, or unauthorized access patterns.
- [DATA_EXPOSURE]: The documentation uses standard placeholders for sensitive information (e.g., 'service-account.json', 'postgresql://...') and describes secure session state management practices.
- [REMOTE_CODE_EXECUTION]: The framework described in the skill supports code execution tools; however, the documentation specifies that these should be used within sandboxed environments like Vertex AI Code Interpreter.
- [INDIRECT_PROMPT_INJECTION]: The safety reference guide specifically addresses the risk of processing untrusted data, providing detailed patterns for input sanitization, PII redaction, and multi-layered defense-in-depth strategies.
Audit Metadata