meeting-scribe

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

• Third-party content exposure detected (high risk: 0.75). The skill ingests user-provided transcript text from the local drop-zone (Read of <basename>.txt in Phase 1/4), which is outsider-authored free text if the user did not author it; that text is then loaded into the agent’s LLM context for classification/summary generation.