ghost-writer

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime LLM context is populated by the user-supplied “Voice DNA Document” (including “Exemplar Passages” and “Ghost Writer Briefing”), which is outsider-authored free text whenever the operating user did not author that document themselves; the LLM reads it directly as input and uses it for drafting.