A structured debugging brief for a failure that resists the obvious fix. Reasoning: high. OpenAI maps complex debugging and deep planning to high. API: reasoning.effort "high". Codex: Astra Extra High is the closest preset; Medium if the failure is routine. Define done in the prompt: Astra is thorough in testing by default, so tell it which results you want to see. Free template — copy and paste into ChatGPT, Claude, or Copilot.