agenticflow-workforce

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs the agent to browse and clone live marketplace templates with commands like "af marketplace list/try" (Live marketplace section) and also references workflow blueprints that perform "Web retrieval + summarize" (e.g., "af workflow init --blueprint summarize-url"), which clearly involves ingesting untrusted, user/third-party web content that can alter deployments and runtime behavior.