agent-workflow

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

• Third-party content exposure detected (high risk: 0.90). The skill includes examples that fetch and ingest arbitrary public web content (e.g., references/tools.md's async fetch_url which uses httpx to fetch any URL and the RequestsToolkit example with allow_dangerous_requests=True) and shows ToolNode and graph templates (assets/graph-template.py and references/tools.md) that execute tool outputs as ToolMessage objects which the agent then reads and incorporates into state, exposing the agent to untrusted third-party content that could enable indirect prompt injection.