Stage 2: Public Target Assignment and Run Settings for the Personal AI Audit

Stage 1 fixed the study before policy-specific targets were assigned. This second registration records the target assignment and random seeds derived from the first eligible public random value after Stage 1. It also records the run window derived from that value's scheduled time and the fixed model versions and access routes. It does not change the sample, hypotheses, prompts, outcomes, tests, stopping rules, or failure rules. No policy from the 54-policy main sample has been used in any model call.