AI Coding Tool Cheat Sheets
Start here for fast, source-backed guides to the AI coding tools developers evaluate most often. The five flagship pages now open with a 60-second cheat sheet before the deeper setup, permissions, security, and team-adoption guidance.
Flagship Cheat Sheets
| Tool | Best starting point when... | Cheat sheet |
|---|---|---|
| Claude Code | you want terminal-first deep repo work and granular permissions | Claude Code |
| OpenAI Codex | you want terminal work, local review, scripting, and AGENTS.md |
OpenAI Codex |
| Cursor | you want an AI-native editor plus terminal agent | Cursor |
| GitHub Copilot | your team is GitHub-centric across IDE, CLI, review, and agents | GitHub Copilot |
| Gemini CLI | you want a Google-backed terminal agent with sandbox and context controls | Gemini CLI |
Not sure which one fits? Start with the AI Coding Tools Comparison.
What Every Flagship Page Gives You
- a 60-second command and configuration reference
- five copy-ready prompts for common developer work
- current installation and session commands
- repository instruction-file guidance
- permissions and safer defaults
- MCP / external-tool guidance
- a team adoption checklist
- verification status and official sources
All Tool Guides
Terminal and CLI First
Editor and IDE First
Model / Integration Coverage
How To Use a Tool Page
- Read the 60-second cheat sheet if the page has one.
- Check
Status,Last verified, andVerification scopebefore copying commands. - Compare the repository instruction and permission model with your current engineering standards.
- Start with a read-only explanation or review task.
- Test interactive editing before headless or broad automation.
- Add MCP, plugins, skills, or external systems only after the base workflow is trusted.
Verification Labels
Documentation verifiedmeans the page was reviewed against official documentation, but the workflow was not fully exercised locally in this repository.Partially verifiedmeans some parts were tested or validated more directly than others.Needs verificationmeans the page is directionally useful but should not be treated as approval-ready without re-checking current official docs.ExperimentalorPlatform-specificmeans rollout assumptions should stay narrow.
A Fair Tool Evaluation
Use the same pilot for each candidate:
- explain an unfamiliar repository without edits
- review a real diff
- plan a small bug fix
- implement the fix interactively
- run project validation
- only then evaluate MCP or non-interactive automation
Record the result with the Tool Evaluation Template.
Verification note: the flagship Claude Code, Codex, Cursor, GitHub Copilot, and Gemini CLI guides were re-checked against official sources on 2026-08-09. Other tool pages retain their own verification dates and should be checked individually.