Add missing test files and update .gitignore to allow test files

This commit addresses the overly restrictive .gitignore pattern that
was excluding all test files (test_*.py), including legitimate pytest
and unittest test suites essential for code quality and CI/CD.

Changes to .gitignore:
- Added negation patterns !**/tests/test_*.py and !**/test_*.py
  to allow proper test files while still blocking temporary scripts
- This enables pytest test suites to be tracked by git

Added test files (17 files):

Telegram Bot Tests (15 files):
- reports-app/telegram-bot/tests/test_auth.py
  Tests for authentication and account linking flow
- reports-app/telegram-bot/tests/test_callbacks.py
  Tests for callback query handlers
- reports-app/telegram-bot/tests/test_formatters.py
  Tests for message formatting utilities
- reports-app/telegram-bot/tests/test_formatters_extended.py
  Extended formatter tests
- reports-app/telegram-bot/tests/test_handlers_menu.py
  Tests for menu handlers
- reports-app/telegram-bot/tests/test_helpers.py
  Tests for helper functions
- reports-app/telegram-bot/tests/test_helpers_extended.py
  Extended helper tests
- reports-app/telegram-bot/tests/test_helpers_real.py
  Real integration tests for helpers
- reports-app/telegram-bot/tests/test_helpers_real_simple.py
  Simplified integration tests
- reports-app/telegram-bot/tests/test_login_flow.py
  Complete login flow integration tests
- reports-app/telegram-bot/tests/test_menus.py
  Menu system tests
- reports-app/telegram-bot/tests/test_session_company.py
  Session and company management tests
- reports-app/telegram-bot/test_claude_integration.py
  Manual integration test (Claude AI)
- reports-app/telegram-bot/test_claude_response.py
  Response formatting test
- reports-app/telegram-bot/test_db.py
  Database operations manual test

Shared Module Tests (2 files):
- shared/auth/test_auth.py
  Authentication system tests
- shared/database/test_pool.py
  Oracle connection pool tests

Security verification:
✅ All test files use mock objects, fixtures, and environment variables
✅ No hardcoded credentials or secrets found
✅ Safe for version control

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

reports-app/telegram-bot/test_claude_response.py

@@ -0,0 +1,86 @@
+"""
+Quick test script to verify Claude's response behavior
+
+⚠️ MANUAL INTEGRATION TEST - Not run by default in CI/CD
+
+This script tests that Claude responds directly without asking for permission
+when using custom tools. Requires Claude API key or claude-code login.
+
+REQUIREMENTS:
+- Claude API key or claude-code login
+- Backend API running on localhost:8001 (for real tool execution)
+
+USAGE:
+  # Run as script
+  python test_claude_response.py
+
+  # Run via pytest (requires -m integration)
+  pytest test_claude_response.py -m integration
+
+NOTE: Test function marked with @pytest.mark.integration
+"""
+import pytest
+import asyncio
+import sys
+import os
+
+# Add parent directory to path
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+
+from app.main import ClaudeAgentWrapper
+
+
+@pytest.mark.integration
+async def test_claude():
+    """Test Claude's response to a simple query"""
+    print("="*60)
+    print("Testing Claude Agent Response")
+    print("="*60)
+
+    # Initialize Claude Agent
+    agent = ClaudeAgentWrapper()
+
+    # Test message
+    messages = [
+        {"role": "user", "content": "Care sunt companiile mele?"}
+    ]
+
+    # Mock JWT token (you'll need to replace this with a real one for actual API calls)
+    jwt_token = "mock_token"
+    telegram_user_id = 999999
+
+    print("\nSending query: 'Care sunt companiile mele?'")
+    print("="*60)
+
+    try:
+        response = await agent.process_message(
+            messages=messages,
+            jwt_token=jwt_token,
+            telegram_user_id=telegram_user_id
+        )
+
+        print("\n📥 CLAUDE RESPONSE:")
+        print("="*60)
+        print(response)
+        print("="*60)
+
+        # Check if response asks for permission
+        permission_keywords = [
+            "permisiune", "permission", "autorizare", "authorization",
+            "pot accesa", "can I access", "să accesez"
+        ]
+
+        asks_permission = any(keyword.lower() in response.lower() for keyword in permission_keywords)
+
+        if asks_permission:
+            print("\n❌ FAIL: Claude is still asking for permission!")
+        else:
+            print("\n✅ PASS: Claude responds directly without asking permission")
+
+    except Exception as e:
+        print(f"\n❌ ERROR: {e}")
+        import traceback
+        traceback.print_exc()
+
+if __name__ == "__main__":
+    asyncio.run(test_claude())