molykit

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow described in SKILL.md is a chat UI/framework (Message.content, PromptInput, SSE parsing) that ingests user-provided chat text, so outsider-authored free text can reach the model via the message/protocol path without needing to select a specific item.