revops-handoffs

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow ingests outsider-authored free text via Breeze AI / call transcripts into LLM context: e.g., the MQL Routing Workflow runs “Breeze AI ‘Summarise Contact’ → populate lead_summary” and the LLM middleware pattern pulls “call summaries (from Fathom/Fireflies)” and “transcripts” to generate handoff documents; these transcripts/summaries originate from third-party customer/prospect conversations (not authored by the operating user).