ChatGPT cheat sheet
Navigation & Search
-
Ctrl + K (or ⌘ + K on Mac) → Quickly search across your chats, files, or projects inside ChatGPT.
-
@ mentions → Use
@to reference files, docs, or people (if you’re in a team workspace). Example:“Summarize @Quarterly_Report and compare it to @Previous_Q_Report.”
-
/commands → Type
/to access shortcuts (e.g.,/new,/summarize,/table,/diagram).
🧠Prompt Crafting
-
Be explicit about the goal.
“Generate a one-page summary of the attached PDF for a technical audience.”
-
Set roles or context.
“You are a senior frontend engineer reviewing my React code.”
-
Use structured instructions.
“List key steps as bullet points → explain each in 2 sentences → end with a quick checklist.”
📎 File & Project Integration
-
You can upload multiple files and then reference them by name (
@filename). -
To analyze spreadsheets, upload
.csvor.xlsxdirectly — then ask:“Plot revenue by region over time.”
-
To draft or iterate on longer text/code, say “use canvas” — it opens a side-by-side editor.
🧰 Efficiency Tips
-
Shift + Enter → Add new lines without sending the message.
-
Ctrl + ↑ / ↓ → Navigate through previous prompts.
-
Cmd/Ctrl + / → Open keyboard shortcut list.
-
Drag-drop files or images right into the chat for instant context.
⚙️ Advanced Prompts
-
Ask for format control:
“Output as Markdown table.”
“Give JSON output only.” -
Iterate precisely:
“Keep the structure identical, but rewrite section 2 for clarity.”
-
Combine reasoning + creativity:
“List 3 design approaches and justify which is most scalable.”
Context Setup
-
Role framing:
“Act as a senior [backend/frontend/devops/etc.] engineer.”
“You’re reviewing my code for maintainability and performance.” -
Context injection:
“Here’s the repo layout: ...”
“This project uses TypeScript + Next.js + MongoDB.”
The clearer your setup, the sharper the answer.
💻 Code Analysis
-
Debugging:
“Find logic bugs in this Python function.”
“Explain why this SQL query runs slowly.” -
Refactoring:
“Rewrite this function to be cleaner and more idiomatic in Go.”
“Make this React component more readable and reusable.” -
Performance tuning:
“Profile likely bottlenecks and propose optimizations.”
“Suggest caching or parallelization strategies.”
🧩 Architecture & Design
-
Design review:
“Sketch a microservice design for this workflow.”
“Compare message queue vs REST API for this system.” -
Trade-off thinking:
“Summarize pros/cons of using Kafka vs RabbitMQ.”
“Which is better here: event-driven or cron-based triggers?” -
System diagram generation:
“Draw a system diagram for this architecture.”
🧮 Data & Queries
-
Database design:
“Normalize this schema.”
“Suggest indexes for this Postgres query.” -
Query debugging:
“Explain why this join duplicates rows.”
“Optimize this aggregation pipeline in MongoDB.” -
Analytics:
“Summarize data from @sales.csv as a Markdown table.”
🧰 DevOps & Automation
-
CI/CD pipelines:
“Generate a GitHub Actions workflow for Node.js test + build.”
-
Docker:
“Create a minimal Dockerfile for FastAPI with caching.”
-
Kubernetes:
“Write a Deployment YAML for this app.”
⚡ Productivity Tips
-
Ctrl + K (or ⌘ + K) → Search all chats, docs, or code discussions.
-
@file → Reference uploaded files or docs directly.
-
/summarize → Auto-condense long logs or PRs.
-
Shift + Enter → Add newlines without sending.
-
Ctrl + / → Show all shortcuts.
🧩 Iteration Prompts
-
Tight feedback loop:
“Simplify without changing logic.”
“Keep function signatures identical but optimize runtime.”
“Add docstrings following Google style.” -
Explain like I’m debugging:
“Walk through what happens line by line.”
“Show how this variable changes over time.”
📈 Bonus: Workflow Patterns
-
Daily code reviews:
“Review @feature_branch.diff for logic and style.”
-
Architecture discussions:
“Outline 3 high-level design options for scaling.”
-
Documentation generation:
“Generate README.md from this project summary.”
No comments:
Post a Comment