feat: immutable entries + full task logging

Entries now immutable once journal is generated:
- Edit/delete returns ENTRY_IMMUTABLE error if journal exists
- Frontend shows lock message and hides delete button
- Delete Journal button to unlock entries

Task logging now stores full JSON:
- request: full JSON request sent to AI provider
- response: full JSON response from AI provider
- prompt: formatted human-readable prompt

Prompt structure:
1. System prompt
2. Previous diary entries (journals)
3. Today's entries

backend/src/services/ai/lmstudio.ts

@@ -1,4 +1,4 @@
-import type { AIProvider, AIProviderConfig } from './provider';
+import type { AIProvider, AIProviderConfig, AIProviderResult } from './provider';
 
 export class LMStudioProvider implements AIProvider {
   provider = 'lmstudio' as const;
@@ -10,7 +10,7 @@ export class LMStudioProvider implements AIProvider {
     this.model = config.model || 'local-model';
   }
 
-  async generate(prompt: string, systemPrompt?: string): Promise<string> {
+  async generate(prompt: string, systemPrompt?: string): Promise<AIProviderResult> {
     const messages: Array<{ role: string; content: string }> = [];
     
     if (systemPrompt) {
@@ -19,26 +19,34 @@ export class LMStudioProvider implements AIProvider {
     
     messages.push({ role: 'user', content: prompt });
 
+    const requestBody = {
+      model: this.model,
+      messages,
+      temperature: 0.7,
+      max_tokens: 2000,
+    };
+
     const response = await fetch(`${this.baseUrl}/chat/completions`, {
       method: 'POST',
       headers: {
         'Content-Type': 'application/json',
       },
-      body: JSON.stringify({
-        model: this.model,
-        messages,
-        temperature: 0.7,
-        max_tokens: 2000,
-      }),
+      body: JSON.stringify(requestBody),
     });
 
+    const responseData = await response.json();
+
     if (!response.ok) {
-      const error = await response.text();
-      throw new Error(`LM Studio API error: ${response.status} ${error}`);
+      throw new Error(`LM Studio API error: ${response.status} ${JSON.stringify(responseData)}`);
     }
 
-    const data = await response.json() as { choices: Array<{ message: { content: string } }> };
-    return data.choices[0]?.message?.content || '';
+    const content = responseData.choices?.[0]?.message?.content || '';
+    
+    return {
+      content,
+      request: requestBody,
+      response: responseData,
+    };
   }
 
   async validate(): Promise<boolean> {