architecture-decision

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text can enter the runtime via the user-provided title/description and any user responses collected through AskUserQuestion widgets (e.g., problem framing, dependencies, GDD linkages, and retrofit missing-section content), which the skill then incorporates into ADR content and passes to subagents for review.