chore(llm): drop unused inputTokens/outputTokens from ChatResult (review #11)

The `chatFn` adapter wired in `runtime-container.ts` returned hardcoded
zero token counts for every call. `LLMProvider.chat()` returns
`Promise<string>` (no usage), so threading real counts here would require
widening that interface across every adapter. Nothing in the
`FirstMentionService` path actually consumed the fields — they only
existed to satisfy the local `ChatResult` shape — so dropping them is
strictly safer than leaving misleading zeros in place. Per-call cost
telemetry continues to flow from `LLMProvider.chat` -> `writeCostEvent`
unchanged.

Updated:
  - `ChatResult` in `first-mention-service.ts` -> `{ text: string }` only,
    with a comment documenting the deliberate decision.
  - `runtime-container.ts` adapter no longer fabricates zero usage.
  - `first-mention-service.test.ts` fixture updated to match.

Author: moralespanitz

src/app/runtime-container.ts

@@ -227,7 +227,12 @@ export function createCoreRuntime(deps: CoreRuntimeDeps): CoreRuntime {
         ],
         { maxTokens },
       );
-      return { text, inputTokens: 0, outputTokens: 0 };
+      // Token usage is intentionally NOT returned here: `LLMProvider.chat`
+      // emits per-call cost telemetry via `writeCostEvent` internally
+      // (see `src/services/llm.ts`). Surfacing zeros at this seam invited
+      // the bug the prior reviewer caught — readers would treat them as
+      // real counts. Drop the field instead until usage is plumbed.
+      return { text };
     },
   );
 

src/services/__tests__/first-mention-service.test.ts

@@ -14,8 +14,6 @@ import { FirstMentionService } from '../first-mention-service.js';
 
 interface ChatResult {
   text: string;
-  inputTokens: number;
-  outputTokens: number;
 }
 
 function makeRepo() {
@@ -28,11 +26,7 @@ function makeRepo() {
 
 function chatReturning(text: string) {
   return vi.fn(
-    async (): Promise<ChatResult> => ({
-      text,
-      inputTokens: 100,
-      outputTokens: 50,
-    }),
+    async (): Promise<ChatResult> => ({ text }),
   );
 }
 

src/services/first-mention-service.ts

@@ -21,10 +21,19 @@ import {
   type FirstMentionEvent,
 } from '../db/repository-first-mentions.js';
 
+/**
+ * Minimal chat-call shape for first-mention extraction.
+ *
+ * Token usage is intentionally not threaded here: cost telemetry for the
+ * underlying LLM call is already emitted from `LLMProvider.chat` (see
+ * `src/services/llm.ts` -> `writeCostEvent`). Reading per-call usage at
+ * this layer would require widening the LLMProvider.chat return type to
+ * include usage and plumbing it through every adapter; until something in
+ * the FirstMentionService path actually consumes it, the extra surface
+ * area would only invite hardcoded zeros that mislead downstream readers.
+ */
 interface ChatResult {
   text: string;
-  inputTokens: number;
-  outputTokens: number;
 }
 
 type ChatFn = (