session-feedback

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

• Third-party content exposure detected (high risk: 0.85). In Write mode, the skill walks the live conversation and emits a Markdown feedback file containing extracted outsider-authored free text (i.e., the operating user’s interlocutor/user-turns), which then becomes readable LLM context in Recall mode when feedback_*.md items are surfaced and listed at runtime.