feat: sync runtime search and schema quality updates from app repo

- port retrieval, validation, and eval improvements relevant to os
- align prompts and dimensions with the flat single-agent model
- replace the old eval suite with the focused core scenarios

Generated with Codex

tests/evals/scenarios/archive/get-dimension.ts

@@ -0,0 +1,15 @@
+import { Scenario } from '../types';
+
+export const scenario: Scenario = {
+  id: 'get-dimension',
+  name: 'Get single dimension',
+  description: 'Fetch detail for a known dimension.',
+  tools: ['getDimension'],
+  input: {
+    message: 'Get details for the dimension "ai".',
+  },
+  expect: {
+    toolsCalledSoft: ['getDimension'],
+    responseContainsSoft: ['ai'],
+  },
+};