crowdin

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly reads and acts on Crowdin project data via Membrane commands (e.g., "membrane action run" and "membrane request" to the Crowdin API in SKILL.md), which will ingest user-generated/untrusted translation strings and tasks from a third-party service that the agent is expected to interpret and can influence subsequent actions.