1254 hand-tested AI prompts across image, video, and music — each with full context, variables, and a sample output. Content is authored in Traditional Chinese; the UI here is English.
讓 Claude Code 掃描你的 repo，產出一份精煉、可驗證、200 行內的 CLAUDE.md 專案記憶檔——含 build/test/lint 指令、目錄結構、慣例、禁止事項，並把不該塞進來的東西改放到 skills 或 path-scoped rules。
用 Plan Mode 的精神，讓 Claude Code 在改任何檔案前先交出一份結構化計畫（影響檔案、步驟、風險、驗收標準），你審查、修正、批准後才執行。避免 AI 一頭熱寫一堆你不要的東西。
讓 Codex 訪談你的專案，再產出一份結構化、不囉嗦的 AGENTS.md，把建置/測試指令、程式風格、commit 規範、禁區一次寫清楚，每次開工自動載入。
不用記正則語法——用中文描述要匹配什麼，AI 給你 regex + 每段在幹嘛 + 測試案例。
用費曼技巧讓 AI 教你：先用大白話 + 比喻講清楚，再點出你最容易誤解的地方，最後給一個自我檢測。
臉書 / Google 廣告測試素材：一次生 10 個不同切角的標題（好奇 / 數字 / 痛點 / 反差 / 急迫…）拿去 A/B。
不露臉短影音的完整腳本：3 秒 hook、中段價值、結尾 CTA，附旁白 + 畫面提示，直接拿去配 AI 影片。
把一篇長文 / 部落格 / 報告，重寫成 6-9 則的 X 或 Threads thread：第一則當 hook、每則一個點、結尾 CTA。
經營個人品牌不用每天想梗——給一個主題，一次產 5 則不同切角（故事 / 觀點 / 教學 / 反差 / 互動）的社群貼文。
把雜亂的會議逐字稿 / 筆記，整理成決議、行動項（負責人 + 期限）、待確認問題，會後 30 秒交差。
By default, helpful = agreeable = burying the real problem in paragraph four. Three one-line instructions reverse it: problem first, force a recommendation, explicitly forbid softening. The difference between reading and fixing.
Most people fill Custom Instructions with a LinkedIn resume, "I'm a software engineer, I like bullet points," and it does nothing. Swap in a description of the conversational relationship you want, and three sentences will completely change how the model behaves.
More than a grammar fix. Purpose-built for persuasive and narrative writing: four priorities (clarity → tone → emotional resonance → structure), an explanation for every change, plus extra suggestions.
Stop picking a role for every decision (financial planner / career coach / relationship coach) — one universal prompt, four steps. The regret line is the core — it forces the model from "it depends" to "actually do this."
Before you log off, ramble into your phone mic for 2 minutes, hand it to Claude, and get back a sorted list: first thing tomorrow morning + ranked to-dos + waiting-on + can-defer. Auto-detects daily vs. weekly mode.
The author wrote this after 30% of their heated replies made them wince the next day — costing a few deals and a friendship. The trick: don't let AI rewrite it, let it show you what the calm version would look like. Once you see it, you usually can't go back to the original.
Upload a 60-minute meeting recording and Gemini surfaces every action item, owner, and timestamp.
給 ChatGPT 產業 + 你的觀點，它用 5 種爆款公式（Contrarian / List / Story / Question / Data）產出 5 篇社群文。
給 ChatGPT 你的內容 topic，它用 10 種爆款標題公式產出 10 個方向。A/B 測試或挑最對的一個。
給 Claude 30 個 SKU 的基本資料（名稱、品類、賣點、材質），一次產出 30 組「標題 + 賣場摘要 + 3 個 bullet feature + 詳情段落」— 小品牌上架效率 10 倍。
客訴信進來、你 5 分鐘要回覆。Claude 用「同理 → 擔責 → 具體補償 → 預防」四步驟框架產出回信 — 客戶覺得被聽見、不會升級到 Google 1 星。
No quiz. No questions. I fed my old chats (things I'd written across years without ever trying to be self-aware) to ChatGPT and ran a prompt that went deeper than any personality test I've tried. 39 traits across 6 lay
I've wasted more hours than I want to admit debugging AI agents that kept going off-script. Switched LLMs, swapped tools, rewrote the logic — turned out the problem was the system prompt the whole time. Too vague, too cr
I applied to a role for three weeks. Recruiter calls, a technical screen, all of it. Then it vanished. The company kept reposting it every 30 days but nobody responded to my final follow-up. Took me an embarrassingly lon