eino

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

• Third-party content exposure detected (high risk: 0.70). The skill's documentation and examples (e.g., references/docs_examples/adk/common/tool/graphtool/examples/2_graph_research/README.md and the SKILL.md Graph/compose examples) describe parallel "Web" searches and multi-source research that ingest public web/KB/local search results which the agent reads and synthesizes as part of its workflow, exposing it to untrusted third‑party content that could carry indirect prompt injections.