# AWR-008-LAB1 — Virtual Cast / Repair Agent gate

This lab is synthetic and offline. It tests whether you can bind a two-axis decision artifact to visible evidence; it does not authorize a technology choice or production write.

1. Pick a shipped fixture and use its published `fixture_ref` from `fixtures/FIXTURE_BINDINGS.json`.
2. Copy the fixture evidence into `evidence/gate_evidence.template.json`; do not invent owners, rights, cost, tests, approval, rollback, or audit evidence.
3. Keep Virtual Cast and Repair Agent decisions independent.
4. For each decision, provide a bounded rationale, an explicit `stop_or_pivot_condition`, and one or more applicable `governance_triggers` (`REVIEW`, `MIGRATION`, or `ESCALATION`). You do not have to use every trigger type.
5. Run `python run_checks.py <your-evidence.json>`.

`CONSISTENT` means only that the visible artifact matches the declared shipped fixture and published consistency rules. The checker never stores or emits an expected final decision.