We use cookies to understand how the site is used and to display ads. Analytics and advertising only run after you accept. You can change your choice anytime. Privacy policy

Skip to content
devvkit
$devvkit learn --librarie github-copilot-guide

GitHub Copilot Guide

[ai][completion][github][vibe-coding]
AI / LLM Tools
Install
# VS Code extension:
# Install "GitHub Copilot" from marketplace

# CLI:
npm install -g @githubnext/github-copilot-cli
# Or: gh extension install github/gh-copilot

# Settings: github.com/settings/copilot

GitHub Copilot is the most widely used AI coding assistant. Its core strength is inline autocomplete: as you type, it suggests the next lines of code. The suggestions are context-aware (reading your open files, imports, and project structure) and improve dramatically with practice.

Copilot Chat (Ctrl+I or the chat panel) lets you ask questions about your codebase, generate code, explain snippets, and fix issues. The `/fix` and `/tests` slash commands handle common tasks. Copilot can also generate PR descriptions from your diff and answer questions about docs repos.

The gh-copilot CLI extends Copilot to the terminal: describe what command you need in plain English and it generates the shell command. This is particularly useful for `docker`, `git`, `kubectl`, and `aws` CLI commands you cannot remember.

Autocomplete

Trigger autocomplete· Get suggestions as you type.
# Copilot suggests automatically as you type.
# Accept: Tab
# Accept word: Ctrl+Right
# Reject: Esc
# Next: Alt+]
# Previous: Alt+[

# Tip: Write a clear function name and parameters
# Copilot will infer the implementation from context
async function fetchUserWithPosts(userId: string) {
  // Copilot suggests the implementation here
}

Chat

Chat: inline· Edit code with AI.
# Select code, press Ctrl+I (Cmd+I on Mac)
# Type your instruction:
"Add input validation"
"Handle the error case"
"Make this more performant"

# Or open chat panel:
# Click the Copilot icon in the sidebar
# Or Ctrl+Shift+I (Cmd+Shift+I)
PR descriptions· Auto-generate PR summary.
# In the GitHub PR page:
# Click "Copilot" button in the PR description box
# Select "Generate with Copilot"
# Copilot reads the diff and writes a summary

# Or use gh CLI:
gh pr create --fill
# Copilot fills the title and description

Slash Commands

Slash commands· Built-in chat commands.
# In Copilot Chat:
/fix       # Fix issues in selected code
/tests     # Generate unit tests
/explain   # Explain the selected code
/doc       # Add documentation comments
/optimize  # Suggest performance improvements
/clear     # Clear chat history

# Example:
# Select a function, type:
/tests # Copilot generates test cases

CLI

Copilot for CLI· Natural language shell commands.
# Install:
gh extension install github/gh-copilot

# Use:
gh copilot suggest "find all files larger than 100MB and compress them"
gh copilot explain "git log --oneline --graph --all"

# Explain: shows what a complex command does
# Suggest: generates commands from description

# Evaluate before running: always review!
gh copilot suggest "find the top 10 CPU hogging processes" -t evaluate

Tips

Context engineering· Get better suggestions.
# Tips for better suggestions:

# 1. Open relevant files: Copilot reads open tabs
# 2. Write good imports first
import { z } from 'zod'
import { prisma } from '@/lib/prisma'

# 3. Use descriptive names
async function calculateMonthlyRevenue(userId: string, year: number, month: number): Promise<number>

# 4. Add type annotations: Copilot uses types
async function syncUserToCRM(user: User): Promise<SyncResult>

# 5. Write a comment describing the goal
// Fetches all orders, filters by date range, and returns aggregated metrics