Skip to content

AI CLI Setup & Usage

Comprehensive guide to setting up and using the AI CLI tools (Claude, Codex, Copilot, Gemini) included in DevOps Images for AI-assisted infrastructure development.

Why AI CLI Tools?

  • Code generation: Generate Infrastructure as Code from natural language
  • Code review: Automated security and best practice reviews
  • Troubleshooting: Debug errors and get solutions
  • Documentation: Auto-generate documentation from code
  • Learning: Get explanations of complex configurations

Available AI CLIs

Tool Provider Best For API Required
claude Anthropic Code review, architecture design, complex reasoning ✅ Anthropic API Key
codex OpenAI Code generation, completion ✅ OpenAI API Key
copilot GitHub IDE integration, inline suggestions ✅ GitHub Copilot subscription
gemini Google Multi-modal tasks, GCP integration ✅ Google AI API Key

Claude CLI Setup

Get API Key

  1. Visit Anthropic Console
  2. Sign up or log in
  3. Navigate to API Keys
  4. Create new API key
  5. Copy the key (starts with sk-ant-)

Configure Claude CLI

docker run -it --rm \
  -v ~/.claude:/root/.claude \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude auth login

Follow the prompts to enter your API key.

docker run -it --rm \
  -v $PWD:/workspace \
  -e ANTHROPIC_API_KEY=sk-ant-... \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Review this code" --file main.tf

Create ~/.claude/config.json: 

{
  "api_key": "sk-ant-..."
}

Then mount it: 

docker run -it --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest

Usage Examples

Code Review

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Review this Terraform code for security issues and best practices" \
    --file terraform/main.tf

Generate Infrastructure Code

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Generate Terraform code to create an AWS VPC with 3 public and 3 private subnets" \
    > vpc.tf

Explain Complex Configuration

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Explain what this Ansible playbook does" \
    --file ansible/deploy.yml

Debug Errors

# Save error output
terraform apply 2>&1 | tee error.log

# Ask Claude to debug
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "This Terraform apply failed. What's wrong and how do I fix it?" \
    --file error.log

Codex CLI Setup

Get API Key

  1. Visit OpenAI Platform
  2. Sign up or log in
  3. Navigate to API Keys
  4. Create new API key
  5. Copy the key (starts with sk-)

Configure Codex CLI

docker run -it --rm \
  -v $PWD:/workspace \
  -e OPENAI_API_KEY=sk-... \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  codex "Generate Python script to list all S3 buckets"

Create ~/.codex/config.json: 

{
  "api_key": "sk-..."
}

Mount it: 

docker run -it --rm \
  -v $PWD:/workspace \
  -v ~/.codex:/root/.codex \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest

Usage Examples

Generate Scripts

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.codex:/root/.codex \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  codex "Create a bash script to backup PostgreSQL database to S3" \
    > backup.sh

Code Completion

# Complete partial code
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.codex:/root/.codex \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  codex --complete \
    --file partial-script.py

GitHub Copilot CLI Setup

Get Access

  1. Subscribe to GitHub Copilot
  2. Install GitHub Copilot CLI extension
  3. Authenticate with GitHub

Configure Copilot CLI

docker run -it --rm \
  -v ~/.copilot:/root/.copilot \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  copilot auth login

Usage Examples

Suggest Commands

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.copilot:/root/.copilot \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  copilot suggest "deploy kubernetes application with helm"

Explain Commands

docker run --rm \
  -v ~/.copilot:/root/.copilot \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  copilot explain "kubectl rollout status deployment/myapp -n production"

Gemini CLI Setup

Get API Key

  1. Visit Google AI Studio
  2. Sign in with Google account
  3. Create API key
  4. Copy the key

Configure Gemini CLI

docker run -it --rm \
  -v $PWD:/workspace \
  -e GOOGLE_API_KEY=... \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  gemini "Generate GCP deployment manager template"

Create ~/.gemini/config.json: 

{
  "api_key": "..."
}

Mount it: 

docker run -it --rm \
  -v $PWD:/workspace \
  -v ~/.gemini:/root/.gemini \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest

Usage Examples

GCP-Specific Tasks

docker run --rm \
  -v $PWD:/workspace \
  -v ~/.gemini:/root/.gemini \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  gemini "Create a Cloud Run service with auto-scaling" \
    > cloud-run.yaml

Multi-Modal Analysis

# Analyze architecture diagram
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.gemini:/root/.gemini \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  gemini "Describe this infrastructure architecture" \
    --image architecture.png

Real-World Workflows

Workflow 1: Security-First Development

#!/bin/bash
# secure-deploy.sh

# 1. Generate infrastructure code with Claude
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Generate secure AWS EKS cluster with encrypted EBS volumes and VPC endpoints" \
  > eks-cluster.tf

# 2. Review with Claude
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Review this EKS cluster code for security issues" \
    --file eks-cluster.tf \
  > security-review.md

# 3. Scan with Trivy
docker run --rm \
  -v $PWD:/workspace \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  trivy config eks-cluster.tf

# 4. Deploy
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.aws:/root/.aws \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  sh -c "terraform init && terraform apply"

Workflow 2: AI-Assisted Troubleshooting

#!/bin/bash
# ai-troubleshoot.sh

# Capture error
kubectl apply -f deployment.yaml 2>&1 | tee error.log

# Get AI help
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "This Kubernetes deployment failed. Analyze the error and provide a fix." \
    --file error.log \
    --file deployment.yaml \
  > solution.md

cat solution.md

Workflow 3: Documentation Generation

#!/bin/bash
# generate-docs.sh

# Generate README for Terraform module
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Generate comprehensive README.md documentation for this Terraform module" \
    --file main.tf \
    --file variables.tf \
    --file outputs.tf \
  > README.md

Workflow 4: Multi-Cloud Translation

# Translate AWS to GCP
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "Convert this AWS Terraform code to equivalent GCP resources" \
    --file aws/main.tf \
  > gcp/main.tf

Comparison Matrix

When to Use Which AI CLI?

Use Case Claude Codex Copilot Gemini
Security review ✅ Best ⚠️ Good ⚠️ Good ⚠️ Good
Code generation ✅ Best ✅ Best ✅ Best ✅ Best
Architecture design ✅ Best ⚠️ Good ❌ Limited ⚠️ Good
Troubleshooting ✅ Best ⚠️ Good ⚠️ Good ⚠️ Good
Multi-modal (images) ❌ No ❌ No ❌ No ✅ Yes
GCP-specific tasks ⚠️ Good ⚠️ Good ⚠️ Good ✅ Best
Cost $$ $$ $ Free tier

Recommendations:

  • General DevOps: Claude (best reasoning)
  • Quick code snippets: Codex or Copilot
  • GCP workloads: Gemini
  • Visual analysis: Gemini (only one with multi-modal)

Best Practices

Effective AI Usage

  1. Be specific: Detailed prompts get better results
  2. Provide context: Include relevant files with --file
  3. Iterate: Refine prompts based on output
  4. Review output: Always review AI-generated code
  5. Combine tools: Use Trivy/TFLint alongside AI review

Security

  1. Never expose API keys: Use environment variables or mounted config files
  2. Review before deploy: AI-generated code should always be reviewed
  3. Scan AI output: Run security scanners on generated code
  4. Cost awareness: Monitor API usage to avoid unexpected bills
  5. Rotate keys: Regularly rotate API keys

Limitations

  • ❌ AI makes mistakes: Always review output
  • ❌ Not always up-to-date: May suggest deprecated approaches
  • ❌ Context limits: Large files may be truncated
  • ❌ Costs add up: Monitor usage

Cost Management

Track Usage

# Create usage tracking script
cat > track-ai-usage.sh <<'EOF'
#!/bin/bash
DATE=$(date +%Y-%m-%d)
USAGE_LOG="ai-usage-$DATE.log"

echo "$(date): Claude API call" >> $USAGE_LOG
docker run --rm \
  -v $PWD:/workspace \
  -v ~/.claude:/root/.claude \
  -w /workspace \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  claude "$@"
EOF

chmod +x track-ai-usage.sh

Cost Estimates (Approximate)

Provider Model Cost per 1M tokens (input) Cost per 1M tokens (output)
Anthropic Claude 3.5 Sonnet $3 $15
OpenAI GPT-4 Turbo $10 $30
GitHub Copilot $10/month (unlimited) N/A
Google Gemini Pro Free tier, then $0.50 $1.50

Troubleshooting

API key not working

Problem: Authentication error when using AI CLI

Solutions: 1. Verify API key is correct and active 2. Check config file permissions: 

chmod 600 ~/.claude/config.json
3. Ensure container can access mounted config: 
docker run --rm -v ~/.claude:/root/.claude \
  ghcr.io/jinalshah/devops/images/all-devops:latest \
  ls -la /root/.claude

Rate limit exceeded

Problem: Too many API requests

Solutions: 1. Add delays between requests 2. Upgrade to higher tier plan 3. Batch requests when possible 4. Cache responses locally

Context too long

Problem: Input file too large for AI model

Solutions: 1. Split large files into chunks 2. Provide only relevant sections 3. Use summary/extract approach: 

# First, summarize
claude "Summarize the main components of this file" --file large-file.tf

# Then, ask specific questions
claude "Review the VPC configuration" --file large-file.tf

Advanced Integration

Pre-commit Hook with AI Review

.pre-commit-config.yaml:

repos:
  - repo: local
    hooks:
      - id: ai-code-review
        name: AI Code Review
        entry: ./scripts/ai-review.sh
        language: system
        files: \.(tf|yml|yaml)$
        pass_filenames: true

scripts/ai-review.sh:

#!/bin/bash
for file in "$@"; do
  docker run --rm \
    -v $PWD:/workspace \
    -v ~/.claude:/root/.claude \
    -w /workspace \
    ghcr.io/jinalshah/devops/images/all-devops:latest \
    claude "Quick security review of this file" --file "$file"
done

CI/CD Integration

See AI-Assisted DevOps Workflows for complete CI/CD integration examples.

Next Steps