fix(confidence-engine): permit empty plausibleInterpretations for honest zero state

- Zod validator: plausibleInterpretations accepts empty string (zero interpretations) while remaining required and string-typed
- Prompt instruction: SECTION B now explicitly requests "" instead of absence prose when no plausible inputs exist
- Report generation/update paths derive hasPlausibleInterpretations from actual synthesized data rather than hard-coding true
- Tests: updated validation contract, added hasPlausibleInterpretations derivation assertions, updated seam tests to use derived values

lib/graph/investigation-overview-synthesis.js

@@ -28,7 +28,7 @@ export { filterEligibleFindings };
 
 const overviewResponseSchema = z.object({
   understanding: z.string().min(1),
-  plausibleInterpretations: z.string().min(1),
+  plausibleInterpretations: z.string(),
 });
 
 const overviewOutputSchema = {
@@ -210,7 +210,7 @@ ${interpSections.length > 0 ? JSON.stringify(interpSections, null, 2) : "(none)"
 9. Neither field may contain recommendations, decisions, confidence scores, next actions, priorities, or readiness assessments.
 10. This is a FRESH synthesis — do NOT treat any previous overview or Current Understanding as input.
 
-Both fields must be non-empty strings.`;
+understanding must be a non-empty string. When SECTION B has no plausible inputs, return plausibleInterpretations as an empty string ("") — do NOT manufacture an absence statement.`;
 
   return prompt;
 }