chore(inference): disable Call 1 end-to-end; Call 2 stays

overnight-whole-app: Call 1 (vision classification) is OUT.
- POST /api/v1/slabs/{id}/inference/call1 always returns 410 call1_disabled
  (no model chat, no inference_log). Prompt seed/revision APIs kept dormant.
- Review: remove "Suggest with AI" / runInference (inferTaxon).
- GenerateText: remove Classify metadata (Call 1) button; Call 2 generate stays.
- PROMPTS.md banner documents Call 1 OUT / Call 2 STAYS.
- Settings: inference_enabled toggle annotated as Call-1 latch (dormant).
- Backend tests updated to assert 410; Call 2 tests unchanged.

frontend/src/api/slabs.ts

@@ -5,7 +5,7 @@
 //   photos.py       POST/GET           /api/v1/slabs/{id}/photos[.../original|processed]
 //   review.py       POST               /api/v1/slabs/{id}/review
 //   publish.py      POST               /api/v1/slabs/{id}/publish
-//   inference.py    POST               /api/v1/slabs/{id}/inference/call1|call2
+//   inference.py    POST               /api/v1/slabs/{id}/inference/call1 (DISABLED 410) | call2
 //   settings.py     GET/PUT            /api/v1/settings ; POST /api/v1/settings/test-woo|test-inference
 //   taxonomy.py     GET/POST           /api/v1/admin/taxonomy[.../sync]
 //   pricing.py      CRUD               /api/v1/admin/pricing[/{id}]
@@ -332,6 +332,7 @@ export interface Call2Result {
   description: string;
 }
 
+/** Call 1 OUT (overnight-whole-app): always fails with 410 call1_disabled. Prefer manual taxonomy. */
 export function inferTaxon(slabId: string): Promise<Call1Result> {
   return fetch(`/api/v1/slabs/${encodeURIComponent(slabId)}/inference/call1`, {
     method: 'POST',