title: "Tools & Integrations" description: "CLI tools, editor plugins, CI/CD connectors, and third-party services that extend AI coding workflows" section: "Ecosystem" readTime: "10 min"

Tools & Integrations

The AI coding ecosystem extends well beyond the core tools. This page catalogs the integrations, plugins, and CLI companions that make Claude Code, GitHub Copilot, and Cursor more powerful in real development workflows.

CLI Companions

Claude Code + Git Hooks

claude -p runs Claude Code non-interactively and prints the result, and it accepts piped input. That makes it usable from git hooks. Add --bare to skip loading hooks, plugins, MCP servers, and CLAUDE.md so scripted calls start faster.

# .husky/pre-commit: AI security scan of the staged diff
RESULT=$(git diff --cached | claude -p --bare "Review this staged diff for hardcoded secrets and OWASP vulnerabilities.
Reply with exactly 'PASS' or 'BLOCK: <reason>'.")
echo "$RESULT"
case "$RESULT" in PASS*) exit 0 ;; *) exit 1 ;; esac

# .husky/commit-msg: AI commit message improvement
ORIGINAL=$(cat "$1")
IMPROVED=$(claude -p --bare "Rewrite this commit message in conventional commit format.
Return only the message: $ORIGINAL")
[ -n "$IMPROVED" ] && echo "$IMPROVED" > "$1"

GitHub Copilot CLI

The old gh copilot extension (suggest / explain) was retired in October 2025. Its replacement is the standalone GitHub Copilot CLI, an agentic assistant for your terminal:

# Install (Node.js 22+)
npm install -g @github/copilot
 
# or: brew install --cask copilot-cli     (macOS / Linux)
# or: winget install GitHub.Copilot       (Windows)
 
# Start it, then sign in with /login on first launch
copilot

There is no one-to-one replacement for the old one-shot suggest and explain subcommands; you describe the task in a session instead. See GitHub's Copilot CLI install guide.

VS Code Extensions

CI/CD Integrations

Copilot Code Review

Copilot code review is a built-in GitHub feature, not a workflow action. There is no github/copilot-pr-review-action. Instead:

• Manual: assign Copilot as a reviewer on a pull request (Reviewers menu).
• Automatic: a user, repository owner, or organization owner can configure Copilot to review pull requests automatically, and repository rulesets can require it. A PR is reviewed when at least one applicable configuration enables it.
• Policy: your organization must enable the "Copilot code review" policy.

See GitHub's Copilot code review docs for current setup.

GitHub Actions: Claude Code

Anthropic maintains an official action, anthropics/claude-code-action. The fastest setup is to run /install-github-app inside Claude Code in your repository (github.com only). It installs the Claude GitHub App, saves your credential as a repository secret (ANTHROPIC_API_KEY or CLAUDE_CODE_OAUTH_TOKEN), and opens a PR with the workflow files.

With a prompt, the action runs in automation mode on any GitHub event:

name: Claude Security Scan
on:
  pull_request:
    types: [opened, synchronize]
 
jobs:
  security:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - uses: actions/checkout@v6
        with:
          fetch-depth: 1
      - uses: anthropics/claude-code-action@v1
        with:
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          prompt: "Review this pull request's changes for OWASP Top 10 vulnerabilities and hardcoded secrets."
          claude_args: "--max-turns 5"

Without a prompt, the action waits for @claude mentions in issue and PR comments. Anthropic also offers Code Review, automatic PR review without maintaining a workflow file. Keep CLAUDE.md concise (it is read on every run), cap work with --max-turns, and store credentials only as GitHub Secrets.

Pre-commit Framework

# .pre-commit-config.yaml
repos:
  - repo: local
    hooks:
      - id: ai-review
        name: AI Code Review
        entry: bash -c 'OUT=$(git diff --cached | claude -p --bare "Flag only critical bugs and security issues. Reply PASS if none."); echo "$OUT"; case "$OUT" in PASS*) exit 0 ;; *) exit 1 ;; esac'
        language: system
        pass_filenames: false

Testing Integrations

Vitest + AI Coverage Analysis

# Generate coverage report, ask AI to analyze gaps
npx vitest run --coverage --coverage.reporter=json-summary
cat coverage/coverage-summary.json | claude -p "Identify the 5 most critical gaps
in test coverage based on this report. Suggest specific test cases for each."

Playwright MCP

Microsoft's Playwright MCP server lets Claude Code drive a real browser for end-to-end checks:

claude mcp add playwright -- npx -y @playwright/mcp@latest

claude "Use Playwright to test the checkout flow on localhost:3000.
Navigate to a product page, add to cart, proceed to checkout,
verify the order summary shows the correct total.
Take a screenshot at each major step and flag any failures."

Claude Code also has a built-in Chrome integration (claude --chrome) for testing web apps in your own browser. See MCP Servers for the config-file form.

Documentation Generation

TypeDoc + AI Enhancement

# Find weak JSDoc in the source, then regenerate the docs
claude "Find exported functions in src/lib/ whose JSDoc is missing or just an
auto-generated stub, and write accurate descriptions, params, and examples."
 
npx typedoc --out docs/api src/

Edit the JSDoc in source rather than the generated HTML, which is overwritten on the next build.

Storybook + Copilot Story Generation

In VS Code, ask Copilot (agent mode) with Storybook open:

Generate Storybook stories for all components in src/components/
that don't have a .stories.tsx file. Include default, loading, and error states.

Productivity Tools

Warp Terminal

Warp has built-in AI command suggestions that combine well with Claude Code. Per Warp's docs, press Ctrl-backtick or type # at the command line to describe a command in plain English. AI features are opt-in and can be turned off in Settings > AI. Shortcuts change between releases, so check Warp's docs for your version.

Claude Code in Your Editor

Claude Code also runs inside VS Code and JetBrains IDEs, the desktop app, on the web, in Slack, and in CI, so you often don't need a separate launcher. Check the Claude Code docs for the current integration list.

Monitoring AI Code Quality

Track the output quality of your AI tools over time: