brainstorm

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

• Third-party content exposure detected (medium risk: 0.30). The skill takes an outsider-provided argument/topic from the ongoing conversation and uses it directly in the runtime prompts to six subagents (“brainstorming about: {topic — two or three sentences of context}”), so free text can be ingested as the topic/context without selecting a specific trusted item.