Add detailed logging for prompts and input data in agent executor to improve debugging and traceability. This includes user text, guidelines count, and conversation context before sending requests to the LLM.

src/agent/agent-executor.ts

@@ -167,6 +167,23 @@ Tạo response JSON (CHỈ JSON thuần, không có markdown):
   }
 }`;
 
+      // Log prompt and input data before sending to LLM
+      logger.info('='.repeat(80));
+      logger.info('[AGENT] PROMPT SENT TO LLM (Disease Info Query):');
+      logger.info(prompt);
+      logger.info('='.repeat(80));
+      logger.info('[AGENT] INPUT DATA SUMMARY (Disease Info Query):');
+      logger.info(`- User text: "${userText}"`);
+      logger.info(`- Guidelines count: ${guidelines.length}`);
+      if (guidelines.length > 0) {
+        guidelines.forEach((g, i) => {
+          const content = typeof g === 'string' ? g : (g.content || g.snippet || JSON.stringify(g));
+          logger.info(`  ${i + 1}. Preview: ${content.substring(0, 200)}...`);
+        });
+      }
+      logger.info(`- Conversation context: ${conversationContext ? 'Yes' : 'No'}`);
+      logger.info('='.repeat(80));
+
       const generations = await this.llm._generate([prompt]);
       const response = generations.generations[0][0].text;
 
@@ -452,6 +469,32 @@ Hãy tạo response JSON với format sau (CHỈ JSON thuần, không có markdo
   }
 }`;
 
+    // Log prompt and input data before sending to LLM
+    logger.info('='.repeat(80));
+    logger.info('[AGENT] PROMPT SENT TO LLM:');
+    logger.info(prompt);
+    logger.info('='.repeat(80));
+    logger.info('[AGENT] INPUT DATA SUMMARY:');
+    logger.info(`- User text: "${userText}"`);
+    logger.info(`- CV results count: ${cvResult.top_conditions.length}`);
+    if (cvResult.top_conditions.length > 0) {
+      cvResult.top_conditions.forEach((c: any, i: number) => {
+        logger.info(`  ${i + 1}. ${c.name}: ${(c.prob * 100).toFixed(1)}%`);
+      });
+    }
+    logger.info(`- Triage level: ${triageResult.triage}`);
+    logger.info(`- Triage reasoning: ${triageResult.reasoning || 'N/A'}`);
+    logger.info(`- Red flags: ${triageResult.red_flags?.join(', ') || 'None'}`);
+    logger.info(`- Guidelines count: ${guidelines.length}`);
+    if (guidelines.length > 0) {
+      guidelines.forEach((g, i) => {
+        const content = typeof g === 'string' ? g : (g.content || g.snippet || JSON.stringify(g));
+        logger.info(`  ${i + 1}. Preview: ${content.substring(0, 200)}...`);
+      });
+    }
+    logger.info(`- Conversation context: ${conversationContext ? 'Yes' : 'No'}`);
+    logger.info('='.repeat(80));
+
     const generations = await this.llm._generate([prompt]);
     const response = generations.generations[0][0].text;