assessment-architect

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required runtime workflow ingests professor- or user-provided text to generate assessment content (e.g., exam mode uses the passport assessment entry and standalone inputs in Phase 0/1, then agents draft items from that context), so outsider-authored free text can reach the LLM without a specific item selection step.