refactor(ultrawork): remove thinking config injection from model override

Delegate thinking config control to the provider layer rather than
injecting it manually in ultrawork model override.

🤖 Generated with assistance of [OhMyOpenCode](https://github.com/code-yeongyu/oh-my-opencode)

src/plugin/ultrawork-model-override.ts

@@ -8,7 +8,6 @@ import { scheduleDeferredModelOverride } from "./ultrawork-db-model-override"
 const CODE_BLOCK = /```[\s\S]*?```/g
 const INLINE_CODE = /`[^`]+`/g
 const ULTRAWORK_PATTERN = /\b(ultrawork|ulw)\b/i
-const ULTRAWORK_THINKING_CONFIG = { type: "enabled", budgetTokens: 16000 } as const
 
 export function detectUltrawork(text: string): boolean {
   const clean = text.replace(CODE_BLOCK, "").replace(INLINE_CODE, "")
@@ -118,7 +117,6 @@ export function applyUltraworkModelOverrideOnMessage(
   if (!override.providerID || !override.modelID) {
     if (override.variant) {
       output.message["variant"] = override.variant
-      output.message["thinking"] = { ...ULTRAWORK_THINKING_CONFIG }
     }
     return
   }
@@ -135,7 +133,6 @@ export function applyUltraworkModelOverrideOnMessage(
     output.message.model = targetModel
     if (override.variant) {
       output.message["variant"] = override.variant
-      output.message["thinking"] = { ...ULTRAWORK_THINKING_CONFIG }
     }
     return
   }