feat(confidence-engine): expose initial reconstruction evidence

docs/current-handoff.md

@@ -19,6 +19,13 @@
 - Targeted deterministic suite passed; live v0.5 semantic/reliability validation remains pending.
 - 502 structured-output reliability remains a separate unresolved issue.
 
+## Initial reconstruction observability
+
+- Successful `/api/cases/start` responses now expose the validated initial `reconstruction`, the exact object used by `buildInitialGraph()`.
+- Semantic experiments can compare `reconstruction.relationships` directly with projected `e-rel-*` SituationGraph edges; this is observability only and does not change reasoning semantics.
+- Owning deterministic tests pass. Ten broader orchestrator failures were observed and classified as unrelated to this increment, but were not proven pre-existing because non-mutating baseline execution was blocked by the existing Vitest apparatus.
+- No live validation has been rerun against this observability boundary. Next restart point: exactly one production-default `/api/cases/start` call using the fixed manufacturing scenario, comparing reconstruction relationship E directly with its projected graph edge.
+
 ## Current product architecture
 
 Three distinct routes, not a single page:

lib/graph/orchestrator.js

@@ -496,6 +496,7 @@ export async function startCase(body, dependencies = {}) {
   return {
     success: true,
     summary: analysis.reconstruction?.summary ?? null,
+    reconstruction: analysis.reconstruction,
     situationGraph,
     selectedQuestion,
     diagnostics: buildDiagnostics({

tests/app/api/cases-start-route.test.js

@@ -33,9 +33,33 @@ describe("app/api/cases/start route", () => {
   });
 
   it("returns 200 on success", async () => {
+    const reconstruction = {
+      summary: "Validated reconstruction summary",
+      relationships: [
+        {
+          id: "r1",
+          fromId: "u-intervention",
+          toId: "u-problem",
+          relationship: "depends_on",
+          description: "The intervention depends on the unresolved problem.",
+          confidence: "high",
+        },
+      ],
+    };
     mockStartCase.mockResolvedValue({
       success: true,
-      situationGraph: { nodes: [{ id: "n1" }], edges: [] },
+      reconstruction,
+      situationGraph: {
+        nodes: [{ id: "n1" }],
+        edges: [
+          {
+            id: "e-unrelated",
+            fromNodeId: "n1",
+            toNodeId: "n1",
+            relationship: "supports",
+          },
+        ],
+      },
       selectedQuestion: null,
       diagnostics: {},
     });
@@ -50,6 +74,10 @@ describe("app/api/cases/start route", () => {
     );
 
     expect(response.status).toBe(200);
+    await expect(response.json()).resolves.toMatchObject({
+      success: true,
+      reconstruction,
+    });
   });
 
   it("returns 400 for invalid request input", async () => {

tests/graph/orchestrator.test.js

@@ -371,12 +371,14 @@ describe("lib/graph/orchestrator startCase", () => {
   });
 
   it("builds a valid graph on successful analysis", async () => {
-    mockAnalyseScenario.mockResolvedValue(makeAnalysisResult());
+    const analysis = makeAnalysisResult();
+    mockAnalyseScenario.mockResolvedValue(analysis);
     const { startCase } = await import("@/lib/graph/orchestrator.js");
 
     const result = await startCase({ scenario: "Scenario text" });
 
     expect(result.success).toBe(true);
+    expect(result.reconstruction).toBe(analysis.reconstruction);
     expect(result.situationGraph.centralStatement).toBe("Scenario text");
     expect(result.situationGraph.currentSummary).toContain("Nodes:");
     expect(result.diagnostics).toMatchObject({