figma-kiwi-protocol-decode

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill includes explicit instructions to "steal cookies" and to open a WebSocket to Figma's multiplayer endpoint to apply live mutations, which enables credential theft and unauthorized remote modification of Figma files.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s read workflow ingests binary scenegraph/SVG/CSS by intercepting Figma WebSocket frames via Chrome DevTools Protocol (CDP) and decoding them, where the underlying content originates from the user’s Figma file and therefore can include outsider-authored text (e.g., Figma comments/text) without needing the agent to pre-select a specific item.