bilalmk__todo_correct/FIX_SUMMARY_500_ERROR.md

Fix Summary: 500 Error When Creating Tasks with Advanced Fields

Problem

When testing: "I need to submit the quarterly report to my manager by January 20th at 2pm, this is very urgent and important"

Error:

HTTP Request: POST http://localhost:8001/mcp "HTTP/1.1 500 Internal Server Error"
Tool todo_add_task failed: Server error '500 Internal Server Error'

But the AI WAS working correctly:

• ✅ Extracted priority: "high" (from "urgent and important")
• ✅ Extracted due_date: "2026-01-20T14:00:00Z" (from "January 20th at 2pm")
• ✅ Extracted title: "Submit the quarterly report to my manager"

The problem was in the MCP tool response formatting, not the AI parsing!

---

Root Cause

The format_task_result() function in mcp_server/src/todo_mcp/utils/responses.py was only returning 3 fields:

{
    "task_id": 42,
    "status": "created",
    "title": "Buy groceries"
}

It was missing:

• ❌ priority - causing JSON serialization to fail when trying to serialize Task object
• ❌ due_date - datetime object not converted to ISO string
• ❌ reminder_at - datetime object not converted to ISO string
• ❌ recurrence_pattern - not included
• ❌ recurrence_config - not included
• ❌ description, completed, created_at, updated_at - basic fields also missing

When json.dumps() tried to serialize the Task model with datetime objects directly, it failed with a 500 error because datetime objects aren't JSON serializable without conversion.

---

Changes Applied

1. Updated format_task_result() (responses.py:12-56)

Before:

def format_task_result(task, status):
    return json.dumps({
        "task_id": task.id,
        "status": status,
        "title": task.title,
    })

After:

def format_task_result(task, status):
    result = {
        "task_id": task.id,
        "status": status,
        "title": task.title,
        "description": task.description,
        "completed": task.completed,
        "created_at": task.created_at.isoformat() if isinstance(task.created_at, datetime) else task.created_at,
        "updated_at": task.updated_at.isoformat() if isinstance(task.updated_at, datetime) else task.updated_at,
        # Advanced fields (Phase V)
        "priority": task.priority,
        "due_date": task.due_date.isoformat() if isinstance(task.due_date, datetime) and task.due_date else task.due_date,
        "reminder_at": task.reminder_at.isoformat() if isinstance(task.reminder_at, datetime) and task.reminder_at else task.reminder_at,
        "recurrence_pattern": task.recurrence_pattern,
        "recurrence_config": task.recurrence_config,
    }
    return json.dumps(result, indent=2)

2. Updated format_task_list() (responses.py:84-121)

Added same advanced fields to task list formatting for consistency.

---

What's Fixed

✅ Datetime Serialization: All datetime fields (due_date, reminder_at, created_at, updated_at) are now properly converted to ISO 8601 strings before JSON serialization

✅ Complete Task Object: Response now includes ALL task fields, not just title

✅ Null Safety: Advanced fields return null if not set (proper JSON null, not Python None)

✅ API Consistency: All MCP tool responses now return complete task objects

---

Testing Instructions

1. Restart MCP Server (if needed)

The MCP server should have auto-reloaded when you saved the changes. Verify:

# Check MCP server is running
curl http://localhost:8001/health

# Should return: {"status": "healthy", ...}

2. Test Case 1: High Priority + Due Date

curl -N -X POST http://localhost:8000/api/chatkit/chat \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d '{"message": "I need to submit the quarterly report to my manager by January 20th at 2pm, this is very urgent and important"}'

Expected Response (SUCCESS):

{
  "type": "message",
  "content": "✅ Created task #X: 'Submit the quarterly report to my manager' (Priority: High, Due: Jan 20, 2026 at 2:00 PM)",
  "tool_results": [{
    "tool": "todo_add_task",
    "arguments": {
      "title": "Submit the quarterly report to my manager",
      "priority": "high",
      "due_date": "2026-01-20T14:00:00Z"
    },
    "result": {
      "task_id": 20,
      "status": "created",
      "title": "Submit the quarterly report to my manager",
      "description": null,
      "completed": false,
      "priority": "high",
      "due_date": "2026-01-20T14:00:00+00:00",
      "reminder_at": null,
      "recurrence_pattern": null,
      "recurrence_config": null,
      "created_at": "2026-01-14T19:45:00+00:00",
      "updated_at": "2026-01-14T19:45:00+00:00"
    }
  }]
}

3. Test Case 2: Weekly Recurring

curl -N -X POST http://localhost:8000/api/chatkit/chat \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d '{"message": "Remind me to water the office plants every Monday morning, not urgent but needs to be done weekly"}'

Expected Response:

{
  "result": {
    "task_id": 21,
    "status": "created",
    "title": "Water the office plants",
    "priority": "low",
    "recurrence_pattern": "weekly",
    ...
  }
}

4. Verify Database

Check that the task was saved with all fields:

# Connect to your database and query
SELECT id, title, priority, due_date, recurrence_pattern
FROM tasks
WHERE deleted_at IS NULL
ORDER BY id DESC
LIMIT 5;

You should see tasks with populated priority and due_date fields!

---

What Should Work Now

✅ Priority Extraction: "urgent", "important" → priority="high" ✅ Due Date Parsing: "January 20th at 2pm" → due_date="2026-01-20T14:00:00Z" ✅ Recurrence Detection: "every Monday", "weekly" → recurrence_pattern="weekly" ✅ Title Cleaning: Removes temporal/priority keywords from title ✅ Complete Responses: All tool responses include all task fields ✅ Proper JSON: Datetime objects converted to ISO 8601 strings

---

Debugging Tips

If you still get errors:

1. Check MCP Server Logs:

   # Look for any errors in the MCP server output
   # It should show successful tool calls now
   

2. Verify Tool Schema:

   # Check that advanced fields are in the tool schema
   curl -s -X POST http://localhost:8001/mcp \
     -H "Content-Type: application/json" \
     -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"0.1.0","capabilities":{},"clientInfo":{"name":"test","version":"1.0"}}}'
   
   curl -s -X POST http://localhost:8001/mcp \
     -H "Content-Type: application/json" \
     -d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}' | grep -A 5 "priority"
   

3. Test Pydantic Validation:

   python3 test_pydantic_datetime.py
   # Should show SUCCESS for both test cases
   

4. Check Backend Logs: Look for the tool call arguments and result to verify the flow

---

Summary

The fix was in 3 locations:

1. ✅ tools_registry.py - Added advanced fields to tool schemas (so AI knows about them)
2. ✅ models/inputs.py - Already had datetime fields (Pydantic parses correctly)
3. ✅ utils/responses.py - THIS WAS THE BUG - Now includes all fields and serializes datetimes

The 500 error was caused by incomplete response formatting, not AI parsing. The AI was working perfectly - it just couldn't return the result properly! 🎉