Make the governed RAG evidence path executable end to end

Separate local parsing from model indexing, bind review decisions to immutable manifests, persist vectors behind active profiles, and expose retrieval, chat, evaluation, and document workflows through the React workbench.

Constraint: Live Bailian authentication currently fails for all three configured capabilities

Rejected: Direct upload-to-embedding flow | bypasses local review and manifest binding

Confidence: high

Scope-risk: broad

Directive: Keep private-data deployment blocked until authentication, RBAC, and separate database roles land

Tested: make verify; fresh and replay Docker document smoke; worker recovery smoke; frozen synthetic evaluation; migration 0003-0004 roundtrip

Not-tested: Successful live Bailian calls, OCR, real multi-user authorization

backend/tests/unit/test_problem_validation.py

@@ -0,0 +1,38 @@
+from __future__ import annotations
+
+import httpx
+import pytest
+
+from app.api.v1.retrieval import get_retrieval_service
+from app.main import create_app
+
+
+@pytest.mark.asyncio
+async def test_request_validation_is_problem_json_without_rejected_value() -> None:
+    secret_shaped_value = "sk-" + "A" * 24
+    app = create_app()
+    app.dependency_overrides[get_retrieval_service] = lambda: object()
+    async with httpx.AsyncClient(
+        transport=httpx.ASGITransport(app=app),
+        base_url="http://test",
+    ) as client:
+        response = await client.post(
+            "/api/v1/retrieval/search",
+            json={
+                "knowledge_base_id": secret_shaped_value,
+                "query": "铜矿",
+                "access_scope_id": secret_shaped_value,
+            },
+        )
+
+    assert response.status_code == 422
+    assert response.headers["content-type"].startswith("application/problem+json")
+    payload = response.json()
+    assert payload["code"] == "REQUEST_VALIDATION_FAILED"
+    assert payload["status"] == 422
+    assert payload["trace_id"] == response.headers["x-request-id"]
+    assert {item["field"] for item in payload["field_errors"]} == {
+        "knowledge_base_id",
+        "access_scope_id",
+    }
+    assert secret_shaped_value not in response.text