Posts about “secret ChatGPT codes” spread fast. The slash versions are usually just compact prompts. ChatGPT does not unlock a hidden mode when you type /eli5. It simply treats that text as an instruction: explain this simply.
That still helps. Short, consistent labels save time if you reuse them.
12 shortcuts worth using
- /human — write like a real person, less stiff
- /expert — specialist-level answer
- /eli5 — explain simply
- /brief — shortest useful answer
- /rewrite — clearer and tighter
- /outline — structure first
- /critic — find weak points
- /teacher — teach step by step
- /research — organized findings
- /brainstorm — more ideas fast
- /email — short professional message
- /improve — make this prompt better
How to use them
/eli5 What is prompt engineering?
Give one example a beginner can copy.
/rewrite Keep the same meaning. Cut fluff. Make this sound like a human wrote it.
Text:
[PASTE]
/outline Then write a short draft.
Topic: [TOPIC]
Audience: [WHO]
Goal: [what they should do next]
Writing shortcuts
- Draft a blog outline
- Rewrite for clarity
- Short email that gets a reply
- Turn messy notes into clean text
- Make it less generic
- Keep the facts, cut the fluff
- Change tone without changing meaning
- Give 3 title options
Time-saving jobs
- Meeting notes → action list
- Daily plan for the next 4 hours
- Research brief under 400 words
- Compare two options clearly
- Debug with context
- Summarize a long page
- Practice questions for studying
- Role + task + constraints + format
The framework that still matters most
You are a [ROLE].
Task: [WHAT TO DO]
Constraints:
- Audience: [WHO]
- Avoid: [fluff / jargon / extra claims]
Format:
- [bullets / short paragraphs / table]
- Length: [target]
Shortcuts are optional. Clear role, task, limits, and format still produce the most reliable results.