Automation & CI/CD

Use Claude Code's headless mode in scripts and CI/CD pipelines

title: "Automation & CI/CD" description: "Use Claude Code's headless mode in scripts and CI/CD pipelines"

Automation & CI/CD ¶

Claude Code supports headless mode for use in scripts, CI/CD pipelines, and automated workflows — no human interaction required.

Basic Usage ¶

Single Execution (-p Mode)¶

# Basic execution
claude -p "Analyze this project's architecture"

# JSON output format
claude -p "List all TODO comments" --output-format json

# Limit execution turns
claude -p "Write unit tests for UserService" --max-turns 5

Restricting Available Tools ¶

# Read-only analysis (no writes allowed)
claude -p "Analyze code quality" --allowedTools Read,Glob,Grep

# Allow read and write (no command execution)
claude -p "Refactor this file" --allowedTools Read,Write,Edit

Skipping Permission Prompts ¶

In secure, isolated environments (e.g., Docker containers or CI environments):

claude -p "Fix all lint errors and commit" --dangerously-skip-permissions

Security warning: --dangerously-skip-permissions skips all permission prompts. Only use this in trusted, isolated environments.

CI/CD Integration Examples ¶

GitHub Actions ¶

name: AI Code Review
on: [pull_request]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:

      - uses: actions/checkout@v4

      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
          node-version: '22'

      - name: Install Claude Code
        run: npm install -g @anthropic-ai/claude-code

      - name: Run AI Code Review
        env:
          ANTHROPIC_BASE_URL: "https://api.qcode.cc/api"
          ANTHROPIC_AUTH_TOKEN: ${{ secrets.QCODE_API_KEY }}
        run: |
          claude -p "Review this PR's code changes, focusing on security and performance issues" \
            --output-format json \
            --max-turns 3 \
            --allowedTools Read,Glob,Grep \
            > review.json

Script Automation ¶

#!/bin/bash
# auto-test.sh — Automatically generate and run tests

export ANTHROPIC_BASE_URL="https://api.qcode.cc/api"
export ANTHROPIC_AUTH_TOKEN="cr_your_api_key"

# Generate tests
claude -p "Generate unit tests for all files under src/services/" \
  --max-turns 10 \
  --allowedTools Read,Write,Glob,Grep,Bash

# Run tests
npm test

Session Resumption ¶

You can maintain context across multiple invocations:

# First call: analysis
claude -p "Analyze the project architecture" --session-id "ci-task-42" --output-format json

# Second call: suggest improvements based on analysis
claude -p "Based on the previous analysis, suggest improvements" --resume --session-id "ci-task-42"

Output Formats ¶

Format	Flag	Description
Plain text	`--output-format text`	Default, human-readable
JSON	`--output-format json`	Structured output, suitable for script parsing
Streaming JSON	`--output-format stream-json`	Real-time streaming output

Environment Variables ¶

Variable	Description
`ANTHROPIC_BASE_URL`	QCode.cc API endpoint
`ANTHROPIC_AUTH_TOKEN`	API key
`CLAUDE_CODE_MAX_TURNS`	Default maximum turns
`CLAUDE_CODE_OUTPUT_FORMAT`	Default output format

Best Practices ¶

Always limit turns: Use --max-turns in CI to prevent infinite execution
Restrict tools: Use --allowedTools to only enable the tools you need
Isolate the environment: Use --dangerously-skip-permissions inside Docker containers
Use JSON output: For automation, prefer --output-format json for easy parsing

Next Steps ¶

See Hooks System for triggering automated actions in CI
See CLI Tips for more command-line usage

← Previous

gpt-image-2 Image Generation and Editing

Hooks System

Automation & CI/CD

Related Documents