fix: retain every direct-model usage call

Multiple model invocations in one task could share a call identifier when
a provider omitted or reused its response ID. Database upserts then replaced
earlier tool-round or summary usage, causing cumulative activity totals to
undercount successful calls.

Each completed direct-model invocation now receives a unique local call ID
while retaining the provider ID as diagnostic context. Tool rounds, final
responses, image calls, and repeated compression summaries are therefore
stored independently.

Release note: 修复连续工具调用或上下文摘要可能覆盖前序模型用量的问题；现在每次成功调用都会分别计入运行记录和累计统计。

src/main/agent/model-runtime.ts

@@ -483,10 +483,13 @@ function createUsageEvent(
   if (!usage.reported) {
     return undefined
   }
+  const localCallId = `model-call:${randomBytes(16).toString('hex')}`
   return {
     requestId,
     type: 'model-usage',
-    callId: (usage.callId ?? requestId).slice(0, 256),
+    callId: usage.callId
+      ? `${localCallId}:${usage.callId}`.slice(0, 256)
+      : localCallId,
     runtime: 'model',
     provider,
     model: (usage.model ?? fallbackModel).slice(0, 500),