social-campaign

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

• Third-party content exposure detected (high risk: 0.75). Phase 2’s “Voice Calibration” reads recent published posts from Typefully via typefully.js drafts:get <id> and uses their full text for analysis, which is outsider-authored free text (content authored by other people on the connected Typefully account) being ingested into the agent’s LLM context.