1. CI/CD Integration

Use Claude Code in your CI/CD pipelines for automated code review, testing, and documentation.

GitHub Actions Example:

name: AI Code Review

on: [pull_request]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

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

      - name: Run Code Review
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
        run: |
          claude -p --output-format json \
            --max-turns 1 \
            "Review the changes in this PR for:
            - Security vulnerabilities
            - Performance issues
            - Code quality
            Output as JSON with 'issues' array" > review.json

      - name: Post Review Comment
        uses: actions/github-script@v7
        with:
          script: |
            const fs = require('fs');
            const review = JSON.parse(fs.readFileSync('review.json', 'utf8'));
            // Process and post review comments

Jenkins Pipeline:

pipeline {
    agent any
    stages {
        stage('AI Review') {
            steps {
                sh '''
                    claude -p --output-format json \
                      --max-turns 3 \
                      "Analyze test coverage and suggest missing tests" \
                      > coverage-analysis.json
                '''
            }
        }
    }
}

2. Script Piping

Process files, logs, and data through Claude for analysis.

Log Analysis:

# Analyze error logs
tail -1000 /var/log/app/error.log | claude -p "summarize these errors and suggest fixes"

# Find patterns in access logs
cat access.log | claude -p "identify suspicious access patterns"

# Analyze git history
git log --oneline -50 | claude -p "summarize recent development activity"

Code Processing:

# Review a specific file
cat src/auth.ts | claude -p "review this authentication code for security issues"

# Generate documentation
cat src/api/*.ts | claude -p "generate API documentation in markdown"

# Find TODOs and prioritize
grep -r "TODO" src/ | claude -p "prioritize these TODOs by importance"

3. Multi-Session Workflows

Manage complex projects with multiple conversation threads.

# Start a feature branch session
claude -r "feature-auth" "let's implement user authentication"

# Later, continue the session
claude -r "feature-auth" "add password reset functionality"

# Fork to try an alternative approach
claude --resume feature-auth --fork-session "try OAuth instead"

# Switch between different feature sessions
claude -r "feature-payments" "continue with Stripe integration"

4. Custom Agent Configuration

Define specialized agents for your team's workflows.

# Save agents config to file
cat > ~/.claude/agents.json << 'EOF'
{
  "reviewer": {
    "description": "Code reviewer for PR reviews",
    "prompt": "Review code for quality, security, and maintainability.",
    "model": "opus"
  },
  "documenter": {
    "description": "Documentation specialist",
    "prompt": "Generate clear, comprehensive documentation.",
    "model": "sonnet"
  },
  "refactorer": {
    "description": "Code refactoring expert",
    "prompt": "Suggest and implement clean code refactoring.",
    "tools": ["Read", "Edit", "Glob"]
  }
}
EOF

# Use agents in session
claude --agents "$(cat ~/.claude/agents.json)" "review the auth module"

5. Batch Processing

Process multiple queries with consistent settings.

# Process multiple files
for file in src/*.ts; do
  echo "Processing $file..."
  claude -p --model haiku "summarize this file: $(cat $file)" >> summaries.md
done

# Batch code review
find src -name "*.py" -exec sh -c '
  echo "## $1" >> review.md
  cat "$1" | claude -p "brief code review" >> review.md
' _ {} \;

# Generate tests for all modules
for module in $(ls src/modules/); do
  claude -p "generate unit tests for src/modules/$module" > "tests/$module.test.ts"
done

6. Security-Conscious Development

Use permission controls for safe operation.

# Read-only security audit
claude --permission-mode plan \
  --tools "Read,Grep,Glob" \
  "audit this codebase for security vulnerabilities"

# Block dangerous commands
claude --disallowedTools "Bash(rm:*)" "Bash(curl:*)" "Bash(wget:*)" \
  "help me clean up this project"

# Restricted automation
claude -p --max-turns 2 \
  --allowedTools "Read" "Glob" \
  "find all hardcoded credentials"

7. JSON API Integration

Use Claude as a programmable API for your tools with jq parsing.

# Get structured analysis
claude -p --output-format json \
  --json-schema '{"type":"object","properties":{"functions":{"type":"array"},"complexity":{"type":"string"}}}' \
  "analyze main.py and return function list with complexity rating"

# Integrate with jq for processing
claude -p --output-format json "list all API endpoints" | jq '.endpoints[]'

# Use in scripts
RESULT=$(claude -p --output-format json "is this code secure? answer with {secure: boolean, issues: []}" < code.py)
if echo "$RESULT" | jq -e '.secure == false' > /dev/null; then
  echo "Security issues found!"
  echo "$RESULT" | jq '.issues[]'
fi

jq Parsing Examples

Parse and process Claude's JSON output using jq:

# Extract specific fields
claude -p --output-format json "analyze this code" | jq '.result'

# Filter array elements
claude -p --output-format json "list issues" | jq -r '.issues[] | select(.severity=="high")'

# Extract multiple fields
claude -p --output-format json "describe the project" | jq -r '.{name, version, description}'

# Convert to CSV
claude -p --output-format json "list functions" | jq -r '.functions[] | [.name, .lineCount] | @csv'

# Conditional processing
claude -p --output-format json "check security" | jq 'if .vulnerabilities | length > 0 then "UNSAFE" else "SAFE" end'

# Extract nested values
claude -p --output-format json "analyze performance" | jq '.metrics.cpu.usage'

# Process entire array
claude -p --output-format json "find todos" | jq '.todos | length'

# Transform output
claude -p --output-format json "list improvements" | jq 'map({title: .title, priority: .priority})'

Overview