M30: record production recruiter acceptance

PROJECT_STATE.md

@@ -2672,3 +2672,28 @@ evidence yet."
 - Exact next action: commit and push M29, create and verify a live database backup,
   deploy the committed archive to Unraid, then run live health, migration, browser and
   acceptance checks before recording the final production evidence.
+
+## M30 — recruiter polish production acceptance (2026-08-10)
+
+- Pushed M29 to Gitea `master` as `29325b6`, created and verified pre-deployment backup
+  `mobilityops-20260810T161659Z.dump`, and deployed that exact source archive while
+  preserving `.env`, volumes and the existing central n8n. Deployment revision is
+  `29325b6c2775806080dc40e1988634d6ceb300a0`; Alembic is `b913a72e8c14 (head)`.
+- API, web, PostgreSQL, scheduled backup, Prometheus and Grafana are healthy/current.
+  Live readiness is `ready`; post-deployment logs contain no traceback or critical error.
+- The initial production browser run exposed two stale assertions that expected raw
+  RAGcore filenames. Fleet Ops now intentionally presents localized source titles. The
+  tests were tightened around localized titles plus language-specific evidence; the
+  focused live matrix passed 4/4 and the complete live suite then passed **150/150 in
+  4.2 minutes**.
+- Final live visual audit covered login, Highlights and Engineering. Desktop overflow is
+  zero and the 360/390 px layouts are automated. The integration story is evidence-based:
+  n8n operational (4/4 workflows, 19 successful, zero pending/unexpected failures),
+  RAGcore operational (11 managed sources, provider index count honestly unknown), and
+  MCP labelled not connected because the final reset clears recent tool-call evidence.
+- Final reset restored **2 users / 180 customers / 50 vehicles / 254 bookings / 75
+  inspections / 40 maintenance records / 33 quality issues / 20 workflow runs** with
+  all five scenarios ready. Live evidence screenshots and the canonical acceptance
+  summary were refreshed in `artifacts/evidence/`.
+- Exact next action: commit/push this evidence-only hand-off update and refresh the
+  server's source archive/revision marker; no runtime rebuild or database change is needed.

artifacts/evidence/final-summary.md

@@ -1,83 +1,73 @@
-# MobilityOps — final acceptance evidence
+# Fleet Ops — final acceptance evidence
 
 ## Accepted implementation
 
-- Implementation under test: `5dda574` (`master`). The final evidence commit contains
+- Runtime commit: `29325b6c2775806080dc40e1988634d6ceb300a0` (`master`).
-  documentation only and does not change the accepted runtime.
+- Live demo: `http://192.168.10.150:1236` — public passwordless synthetic demo.
-- Live URL: `http://192.168.10.150:1236`.
+- Deployment: Unraid `/mnt/user/appdata/mobilityops`, Compose project `mobilityops`.
-- Deployment: Unraid, `/mnt/user/appdata/mobilityops`, Compose project `mobilityops`.
+- Database migration: `b913a72e8c14 (head)`.
-- Data: deterministic, explicitly labelled synthetic demo data.
+- Product name: Fleet Ops; MobilityOps remains the technical repository/deployment ID.
-- Database migration: `f43d829ab610 (head)`.
 
-## Clean acceptance results — 2026-08-10
+## Clean acceptance — 2026-08-10
 
 ```text
-Backend, isolated PostgreSQL: 225 passed, 1 dependency deprecation warning (96.71 s)
+Backend, isolated PostgreSQL: 239 passed, zero warnings
-Playwright, live production:   145 passed (4.4 min)
+Playwright, fresh local build: 150 passed (5.2 min)
+Playwright, live production:   150 passed (4.2 min)
 Ruff:                          clean
-Mypy:                          clean
+Mypy:                          clean across 58 source files
-Frontend TypeScript/Vite:      production build passed
+Frontend lint/build:           passed
-npm audit (prod and full):     0 vulnerabilities
+npm audit, full and prod:      0 vulnerabilities
-Alembic:                       f43d829ab610 (head)
+Alembic:                       b913a72e8c14 (head)
 ```
 
-Backend tests ran through `compose.yaml` plus `compose.test.yaml` in the disposable
+The browser suite covers both roles, protected routes, booking/return/maintenance/user/privacy workflows, five data-quality resolution types, correlated audit, retryable orchestration, grounded knowledge in three languages, the complete guided demo, the recruiter highlights, keyboard behaviour and responsive layouts from 360 to 1440 px.
-`mobilityops-test` project. Its database, network and volume were removed afterward; no
-test process ran inside the live API container.
 
-The single serial Playwright run covers 145 browser scenarios: both roles, protected
+The first production run correctly exposed two stale test assumptions: the redesigned Knowledge Hub shows localized source titles rather than raw file names. The evidence contract was changed to require the localized title and the language-specific source fragment. That focused matrix passed 4/4, followed by the complete 150/150 green production run.
-routes, booking lifecycle and return, maintenance/release, user administration,
-data-quality scanning/assignment/resolution, audit, retryable automation, Knowledge Hub,
-the complete guided demo, keyboard accessibility and responsive layouts from 360 to
-1440 px.
 
-## Live operational evidence
+## Production hand-off state
 
-After acceptance, demo reset restored an internally checked hand-off state:
+The final reset completed at `2026-08-10T16:31:27Z` and restored:
 
 ```text
 users 2 · customers 180 · vehicles 50 · bookings 254 · inspections 75
 maintenance 40 · data-quality issues 33 · workflow runs 20
-scenario_integrity.all_ready = true
+scenario_integrity.all_ready = true (5/5 scenarios)
 ```
 
-Readiness returned `ready` with PostgreSQL `up`; API, database and web containers were
+API readiness is `ready`, PostgreSQL is `up`, and API, web, database, scheduled backup, Prometheus and Grafana are running healthy/current. No traceback or critical error appears in post-deployment API/backup logs.
-healthy. A verified custom-format database backup was written to:
 
-`/mnt/user/appdata/mobilityops/backups/postgres/mobilityops-20260810T111913Z.dump`
+A pre-deployment custom-format backup was created and independently verified by SHA-256 plus `pg_restore --list`:
 
-## External integrations
+`/mnt/user/appdata/mobilityops/backups/postgres/mobilityops-20260810T161659Z.dump`
 
-- **n8n:** MobilityOps uses the existing central server instance, not a second n8n.
+## External integration evidence
-  Dispatch is enabled and operational with 19 successful persisted runs, zero pending,
-  zero unexpected failures and healthy heartbeat evidence for the vehicle-return and
-  scheduled-quality workflows. One failed item is an explicitly labelled synthetic
-  retry scenario and is excluded from incident counts.
-- **RAGcore:** live health returned available/ready for tenant
-  `northstar-mobility-demo`, workspace `mobilityops`, collection
-  `internal-procedures`. A live question returned a grounded answer in 5.1 seconds with
-  three unique, collapsible procedure citations. The UI never substitutes a fabricated
-  answer when evidence is insufficient or the provider is unavailable.
-- **ITWorx MCP Hub:** registration is enabled, the Hub is reachable and MobilityOps
-  reports an audited real `fleet_ops_get_operations_summary` tool call. Tool endpoints
-  are read-only, tenant-bound, client-identity validated, service-authenticated and emit
-  correlation IDs.
 
-## Visual acceptance
+- **Existing central n8n:** configured, dispatch enabled and operational. All four expected workflows have recent healthy heartbeats. The reset state contains 19 successful runs, zero pending, zero unexpected failures and one explicitly synthetic failed-retry scenario.
+- **RAGcore:** reachable and ready for tenant `northstar-mobility-demo`, workspace `mobilityops`, collection `internal-procedures`. Fleet Ops manages 11 authoritative Dutch sources; the latest n8n report records 33 synchronized language documents and zero failures. RAGcore exposes no reliable index-size endpoint, so `document_count` deliberately remains `null`.
+- **ITWorx MCP Hub:** registration is enabled and the Hub is reachable. A reset intentionally clears demo audit history, so the final state has no recent tool-call evidence and the UI labels MCP **not connected** rather than inferring success from configuration. Four tenant-bound service-authenticated endpoints remain contract-tested and read-only.
 
-The production UI was interactively inspected after the automated run. Desktop
+## Recruiter and visual acceptance
-dashboard, mobile data-quality workbench, Integration Management and Knowledge Hub had
-no overlap or horizontal page overflow. Filters, SLA/assignment controls, integration
-states, citations, feedback controls, keyboard landmarks and mobile navigation remained
-visible and operable. Route-level lazy loading keeps the initial JavaScript chunk near
-212 kB instead of the former 572 kB monolith.
 
-## Deliberate scope boundary
+The public entry now offers a direct **90-second Highlights** route. It links three product actions to their engineering proof, while the **Engineering** workspace explains domain ownership, commit-before-orchestration, citation-bound AI, security and explicit scope.
 
-This is the completed locked MobilityOps PoC, not an unrestricted ERP. It intentionally
+Interactive live inspection confirmed:
-does not add accounting, payments, public reservations, CRM, inventory, HR, a second RAG
+
-stack, a separate MCP server or autonomous write tools. Production deployment for real
+- no horizontal overflow on Highlights or Engineering at desktop; automated coverage confirms 360/390 px mobile layouts;
-personal data would still require an organisation identity provider, secrets rotation,
+- clear hierarchy, consistent primary actions and readable proof cards;
-central observability/alerting, retention policy and an operated disaster-recovery
+- truthful live n8n/RAGcore/MCP labels after reset;
-schedule. Those are deployment-governance concerns, not dead or simulated product
+- no dead links, placeholder numbers or unexplained raw data-quality references;
-functions in the accepted scope.
+- complete synthetic-data disclosure on login and in the persistent shell.
+
+Current evidence images are in `artifacts/evidence/screenshots/`:
+
+- `1-login.png`
+- `2-highlights.png`
+- `3-engineering-story.png`
+- `4-dashboard.png`
+- `5-knowledge-evidence.png`
+- `6-highlights-mobile.png`
+
+## Deliberate boundary
+
+This is a completed, production-shaped public demo—not claimed customer adoption and not a general ERP. Accounting, payments, public reservations, CRM, inventory, HR, a second RAG stack, a separate MCP server and autonomous write agents remain intentionally excluded. A real personal-data rollout would additionally require the adopting organisation's identity provider, retention approvals, secrets lifecycle, alert ownership and disaster-recovery governance.

frontend/e2e/demo.spec.ts

@@ -75,7 +75,7 @@ test("five-minute demo script end to end", async ({ page, request }) => {
       .fill("Wat moet ik doen wanneer een voertuig terugkomt met schade?");
     await page.getByRole("button", { name: "Vraag stellen" }).click();
     await expect(page.getByText("Onderbouwd met geciteerde procedures")).toBeVisible({ timeout: 15_000 });
-    const damageSource = page.locator(".source-card", { hasText: /damage-procedure\.md|Procedure schadeafhandeling/ }).first();
+    const damageSource = page.locator(".source-card", { hasText: /zichtbare of gemelde schade/i }).first();
     await expect(damageSource).toBeVisible();
     await expect(damageSource).toContainText(/zichtbare of gemelde schade/i);
   });

frontend/e2e/fleet-ops-correction.spec.ts

@@ -260,7 +260,7 @@ test.describe("knowledge base is grounded in the operator's own language", () =>
       await page.getByRole("button", { name: /^(Vraag stellen|Ask|Demander)$/ }).click();
       const localizedSource = page.locator(".source-card", { hasText: sourceHint }).first();
       await expect(localizedSource).toBeVisible({ timeout: 15_000 });
-      await expect(localizedSource).toContainText(/damage-procedure\.md|Damage handling|schadeafhandeling|gestion des dommages/i);
+      await expect(localizedSource).toContainText(/Schade bij voertuigretour|Vehicle return damage|Dommages au retour du véhicule/i);
     });
   }
 });