kmassets:audit Hardening Follow-ups

Area: solr / kmassets / audit / DX Raised during: Session 2026-07-07 (Sprint 1 1a.9, PR #19) Jira: (add when available) Priority: Low–Medium — non-blocking polish on the working kmassets:audit command

Two follow-ups flagged during the 1a.9 kmassets:audit build. The command works and is verified end-to-end; these are hardening items, not correctness bugs.

1. Confirmation guard for large --fix runs

kmassets:audit --fix reindexes every missing doc. Against a freshly-migrated staging index that is the whole bundle — potentially ~111k writes to the shared staging Solr master. That is legitimate as the deliberate bulk-index, but there is currently no guard against running it unintentionally. Add a confirmation prompt (or a --yes / dry-run --report-only split, or a threshold above which it refuses without --force) so a large --fix is always deliberate.

2. Pass A performance — avoid full entity loads

The missing/stale pass (KmassetAuditor, Drupal→Solr) loads full node entities in batches just to read id + changed — ~1.5 min for 111k on DDEV. A targeted query for nid + changed (skipping full entity hydration) would cut that dramatically. Worth doing if the audit is run frequently or against larger asset types than Images.

Context

Both live in drupal/web/modules/custom/mandala_kmassets_sync/ (the KmassetAuditor service + kmassets:audit Drush command). See kmassets uid identity for the surrounding identity/audit design.