rcs-agent-onboarding

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the workflow routes the user-provided “launch evidence packet” (including URLs, message examples, opt-in/opt-out wording, and support links authored by the user) into a human handoff to Sent via the customer’s free-text email they prepare (“A clean handoff reads like this…”), which the agent/LMM will ingest to compose that handoff note.