Bug Report Format for AI Fixes: Symptom, Repro, Behavior, Environment
Report a bug in this format and the AI's fix hit rate jumps noticeably: give all four parts at once — the raw symptom, a minimal repro, expected behavior, and environment — and force the AI to diagnose first, ask for more information when it's insufficient, and never guess-and-change blindly. Suited to any scenario where you're asking ChatGPT or Claude to fix code. Note: the AI's diagnosis is reasoning, not an actual test run — after the fix, be sure to actually run through the included verification checklist.
【用途】我要你幫我修一個 bug。以下是完整案情、請先診斷再動手、不要看到第一個可疑點就開改。 【症狀】 [症狀描述] (實際發生了什麼、原始錯誤訊息全文照貼、不要自己改寫或節錄) 【最小重現步驟】 [重現步驟] (從乾淨狀態開始、一步一步、寫到走到哪一步壞掉) 【期望行為】 [期望行為] (一切正常的話、應該看到什麼) 【環境】 [環境資訊] (語言與框架版本、作業系統、瀏覽器、套件版本;可能相關的都列) 【相關程式碼】 [相關程式碼] (貼出錯的函式與呼叫它的地方、附檔名與行號) 【已經試過的方法】 [已試方法] (試過什麼、結果如何——避免你重複建議我試過的東西) 【修復要求】 1. 先給診斷:最可能的根因與理由、必須引用我提供的證據、不准腦補我沒描述的行為 2. 資訊不夠就先開口要:列出要我補跑的指令、要貼的輸出、拿到再診斷、禁止瞎猜著改 3. 給修法時只改必要的地方、逐處說明「為什麼這樣改」、標明檔案與位置 4. 最後附「驗證清單」:我要跑什麼指令、看到什麼結果才算真的修好
Swap the variables inside the [ ] brackets for your own content, then paste into ChatGPT / Claude / Gemini.
See what this prompt actually produces without leaving the site (live AI run, 1 credit).
When AI botches a bug fix, it's usually not that the model is dumb — it's that whoever reported the bug just said "it's broken, fix it." The four-part kit assembles all the evidence needed for diagnosis at once: the raw symptom prevents clues from being distorted through paraphrasing, the minimal repro pins down the trigger path, expected behavior defines what "fixed" actually looks like, and environment rules out version-specific issues. "Ask for more information when it's insufficient" is the key safeguard, blocking the model's worst habit — changing a bunch of things that weren't broken based on a guess. Limitation: the AI can't see your machine, and its diagnosis is reasoning, not an actual test — after the fix, be sure to actually run through the verification checklist it provides; if it guesses the wrong direction, paste back the output it asked for, and the second round usually converges on the fix.
[症狀描述]例:按下送出按鈕後畫面白屏、console 顯示 TypeError: Cannot read properties of undefined (reading 'map')
[重現步驟]例:1. npm run dev 啟動 2. 登入測試帳號 3. 進入訂單頁按「送出」→ 白屏
[期望行為]例:按下送出後應顯示成功訊息、並跳轉到訂單列表頁
[環境資訊]例:Next.js 15.1、Node 22、Windows 11、Chrome 138
[相關程式碼]例:貼 app/orders/page.tsx 的 handleSubmit 函式與呼叫端、附行號
[已試方法]填下面的欄位,上方 prompt 會即時替換 [方括號] 內容。填好後按「複製組好的 prompt」直接丟進工具。
【用途】我要你幫我修一個 bug。以下是完整案情、請先診斷再動手、不要看到第一個可疑點就開改。 【症狀】 [症狀描述] (實際發生了什麼、原始錯誤訊息全文照貼、不要自己改寫或節錄) 【最小重現步驟】 [重現步驟] (從乾淨狀態開始、一步一步、寫到走到哪一步壞掉) 【期望行為】 [期望行為] (一切正常的話、應該看到什麼) 【環境】 [環境資訊] (語言與框架版本、作業系統、瀏覽器、套件版本;可能相關的都列) 【相關程式碼】 [相關程式碼] (貼出錯的函式與呼叫它的地方、附檔名與行號) 【已經試過的方法】 [已試方法] (試過什麼、結果如何——避免你重複建議我試過的東西) 【修復要求】 1. 先給診斷:最可能的根因與理由、必須引用我提供的證據、不准腦補我沒描述的行為 2. 資訊不夠就先開口要:列出要我補跑的指令、要貼的輸出、拿到再診斷、禁止瞎猜著改 3. 給修法時只改必要的地方、逐處說明「為什麼這樣改」、標明檔案與位置 4. 最後附「驗證清單」:我要跑什麼指令、看到什麼結果才算真的修好
Suno Engineer's Mindset: 4 Steps to a Song That Doesn't Sound Like AI
A studio engineer's breakdown of Suno's fatal weaknesses (fried vocals, high-frequency artifacts), plus a 4-step DAW workflow and a Suno Studio cleanup prompt.
5 Claude Weekly Workflows That Stuck After 6 Months
Proposal generator / meeting processor / content repurposer / Friday review / shutdown reset — out of 40 I tried, only these 5 survived, each saving 30+ minutes per run.