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/update-node.ts

@@ -0,0 +1,15 @@
+import { Scenario } from '../types';
+
+export const scenario: Scenario = {
+  id: 'update-node',
+  name: 'Update node content',
+  description: 'Create and then update a node in one request.',
+  tools: ['createNode', 'updateNode'],
+  input: {
+    message: 'Create a node titled "Eval: Update Node" with one sentence, then append one more sentence to it.',
+  },
+  expect: {
+    toolsCalledSoft: ['createNode', 'updateNode'],
+    responseContainsSoft: ['updated', 'appended'],
+  },
+};