xquik

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

• Third-party content exposure detected (medium risk: 0.65). SKILL.md documents MCP/REST calls (e.g., tweet search/lookup and explore/schema reads) whose runtime responses include outsider-authored X text (tweets/bios/DM-like content) that the agent is instructed to wrap but still ingests as LLM-readable data; additionally, MCP explore/tool arguments can include schema details and failures but the primary injection surface is the returned X content.