create-figure

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

• Third-party content exposure detected (high risk: 0.80). The skill ingests and acts on untrusted, user-supplied content — e.g., analysis.generate_from_assess reads arbitrary assess JSON to drive figure generation (analysis.py) and the Walkthrough explicitly reports that figure_lab's "promote" copies arbitrary HTML without validation and D3 templates accept window.INJECTED_DATA (d3/gallery and wrap_d3_canvas), meaning third-party/user-generated content can influence rendering and downstream tool actions.