repo-harness-deploy

# repo-harness-deploy Use this command when the user wants a focused deployment and operations configuration check for the repo-local harness. ## Protocol 1. Confirm the target repo path and keep the pass read-only by default. 2. Read `.ai/harness/policy.json` and resolve the optional `operations.deploy_sql` contract before inspecting SQL paths. When absent, use `deploy/sql/` with `ordered4` names; when present, its roots, naming modes, and `invariant_file` are the sole alternate layout authority. 3. Check the tracked deployment surface: - `deploy/` - roots named by `operations.deploy_sql.roots` when configured, otherwise `deploy/sql/` - `deploy/env/.env.example` - release, submission, or runbook files when present 4. Check the private operations boundary: - `_ops/` is ignored - real env files, secrets, provider state, artifacts, logs, and scratch files stay under `_ops/` 5. Run: - `repo-harness run check-deploy-sql-order` 6. Report readiness gaps as concrete missing or misplaced files, not as an app deployment attempt. ## Failure Modes - If `_ops/` contains required private state, report the path class without printing values. - If `operations.deploy_sql` is malformed, stop at the policy error instead of guessing a layout. - If deploy SQL order fails, stop at `check-deploy-sql-order.sh`. - If the user asks to publish or deploy, route to the project release/deploy process instead of this read-only check. ## Boundaries - Read-only by default. - Does...