google-cloud-solution-agentic-ai-data-science-workflow

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow’s only outsider-authored free text exposure is through user-provided requirements/clarifying answers collected in Phase 1 (Steps 1–5), while Phase 2’s documentation fetches target first-party Google Cloud docs (grounding content rather than outsider submissions).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent at runtime to fetch and read external documentation and sample code (e.g., https://docs.cloud.google.com/architecture/agentic-ai-data-science.md.txt and https://github.com/google/adk-samples/tree/main/python/agents/data-science) to "ground" guidance, meaning remote content will directly influence generated prompts/instructions.