jtbd-interview-conductor

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

• Third-party content exposure detected (high risk: 0.75). The required workflow explicitly asks the operating user to “Gather Pre-Interview Context” by collecting “Confluence pages or documentation” and “Slack threads about the decision,” which are outsider-authored free-form text that the agent would ingest into its LLM context via those links/threads at runtime.