fix(confidence-engine): supply overview-specific output schema to generateReconstruction

synthesizeInvestigationOverview() called provider.generateReconstruction(prompt, modelName)
without an output contract. Without a structured-output schema the provider defaults to its
initial-reconstruction format which lacks { understanding, plausibleInterpretations }, causing
'Missing or invalid required fields' validation failure.

Add overviewOutputSchema (JSON-schema matching validateOverviewResponse contract) and pass as
third argument, matching the proven pattern in synthesizeCurrentUnderstanding().

53/53 tests PASS. Co-Authored-By: Claude Code <noreply@anthropic.com>

lib/graph/investigation-overview-synthesis.js

@@ -31,6 +31,16 @@ const overviewResponseSchema = z.object({
   plausibleInterpretations: z.string().min(1),
 });
 
+const overviewOutputSchema = {
+  type: "object",
+  properties: {
+    understanding: { type: "string" },
+    plausibleInterpretations: { type: "string" },
+  },
+  required: ["understanding", "plausibleInterpretations"],
+  additionalProperties: false,
+};
+
 const FORBIDDEN_FIELD_NAMES = new Set([
   "recommendation",
   "decision",
@@ -245,6 +255,7 @@ export async function synthesizeInvestigationOverview(params, deps) {
     providerResult = await provider.generateReconstruction(
       prompt,
       modelName,
+      overviewOutputSchema,
     );
   } catch (err) {
     throw new Error(`Overview synthesis provider call failed: ${err.message}`);