eino-adk
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The architecture for building AI agents described in the skill creates a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the agent context through the
Runner.RunandRunner.Querymethods documented inSKILL.mdandreferences/api-reference.md. - Boundary markers: The examples and documentation do not provide or suggest the use of delimiters or explicit 'ignore' instructions to isolate untrusted input from the system prompt.
- Capability inventory: The framework permits agents to execute arbitrary logic via tools defined with
tool.BaseToolorutils.InferTool(as seen inreferences/examples.md), which may perform file or network operations. - Sanitization: No explicit sanitization or validation of external input is demonstrated before the content is passed to the LLM or tool implementations.
Audit Metadata