liveblocks

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

• Third-party content exposure detected (high risk: 0.80). The skill clearly ingests arbitrary user-generated, third‑party content from Liveblocks rooms (e.g., room.broadcastEvent subscription, useThreads/useStorage/useOthers for chat/comments/presence shown in SKILL.md) which the agent reads and uses to update UI/state, so those remote messages could carry instructions that influence behavior.