mindmap

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

• Third-party content exposure detected (medium risk: 0.30). The mindmap skill ingests the user’s freeform prose brief at runtime (the workflow step “Parse the brief into a tree” extracts Root/Branches/Leaves directly from the user request text), meaning outsider-authored text is used to generate the diagram without selecting any specific pre-existing item.