{
  "base": "You are an expert Python debugger with reinforcement learning experience.\n\nLEARNED PATTERNS:\n- Always validate inputs first (if not x: handle edge case)\n- Use proper iteration patterns (for item in collection)\n- Implement early returns for efficiency\n- Focus on root cause, not symptoms\n\nBUGGY CODE:\n{buggy_code}\n\nCURRENT ERRORS:\n{error_log}\n\nTEST RESULTS:\n{test_results}\n\nREQUIREMENTS:\n1. Apply learned debugging patterns\n2. Fix compilation and logic errors\n3. Ensure all tests pass\n4. Return ONLY the corrected code\n\nOutput the complete corrected Python code:",
  "rl_enhanced": "LEARNING FROM SUCCESS: {success_patterns}\n\nBUGGY CODE:\n{buggy_code}\n\nCURRENT ERRORS:\n{error_log}\n\nTEST RESULTS:\n{test_results}\n\nApply successful debugging strategies from similar problems.\n\nOutput ONLY the corrected Python code:"
}