API Testing Skill (atest)

Claude Code Skill for api-testing (atest) This skill helps you create, manage, and debug API test suites using the api-testing tool.

How to Use This Skill

1. Copy this skill file to your Claude skills directory:

   # On macOS/Linux
   cp claude-skill.md ~/.claude/skills/api-testing.md
   
   # On Windows
   copy claude-skill.md %USERPROFILE%\.claude\skills\api-testing.md
   

2. Restart Claude Code to load the skill

3. Use trigger phrases like:

   • "create an API test suite"
   • "write a test for API"
   • "create grpc test"
   • "API load testing"
   • "mock API server"

What is api-testing (atest)?

api-testing is a comprehensive API testing framework that supports:

• Multi-protocol: HTTP/REST, gRPC, GraphQL
• Load Testing: Duration-based, thread-based, QPS-based
• Mock Server: Create mock APIs from test suites
• Code Generation: Generate test code in Go, Python, Java, JavaScript
• Multiple Reports: Markdown, HTML, PDF, JSON, Prometheus
• Web UI: Built-in interface for test management

Description

You are an expert in the api-testing tool (CLI command: atest), a comprehensive API testing framework written in Go. Help users create test suites in YAML format, run tests via CLI, debug failures, and utilize advanced features like load testing, mocking, and multi-protocol support (HTTP, gRPC, GraphQL).

Trigger Phrases

• "create an API test suite"
• "write a test for API"
• "api-testing test" or "atest test"
• "create grpc test"
• "API load testing"
• "mock API server"
• "convert API tests"
• "debug test suite"

Core Concepts

Test Suite Structure

All test suites use YAML format with these key components:

#!api-testing
name: TestSuiteName
api: https://api.example.com
param:
  key: value
items:
  - name: testCaseName
    request:
      api: /endpoint
      method: GET
      header:
        Authorization: Bearer {{.token}}
      body: |
        {"key": "value"}
    expect:
      statusCode: 200
      verify:
        - data.field == "expected"
    before:
      items:
        - "setupCommand()"
    after:
      items:
        - "teardownCommand()"

Protocol Types

• HTTP/REST (default): Standard REST APIs
• gRPC: gRPC services with proto files
• GraphQL: GraphQL endpoints

Key Features

1. Templating: Uses Sprig template functions with custom additions
2. Data Sharing: Response data from one test available in subsequent tests via data.field
3. Verification: Expression-based verification using expr library
4. Load Testing: Duration-based, thread-based, or QPS-based
5. Mock Server: Create mock APIs from test suites
6. Report Formats: Markdown, HTML, PDF, JSON, Prometheus

Common Commands

# Run a test suite
atest run -p testsuite.yaml

# Run with load testing
atest run -p testsuite.yaml --duration 1m --thread 3 --qps 10

# Run with specific report format
atest run -p testsuite.yaml --report md --report-file report.md

# Generate sample test suite
atest sample

# Start server mode with web UI
atest server --port 7070 --http-port 8080

# Create mock server
atest mock -p testsuite.yaml --port 9090

# Convert tests to JMeter
atest convert -p testsuite.yaml --converter jmeter -t output.jmx

# Install as service
atest service install

Pre-Test Execution Checklist

Before running test suites, always verify the target service is accessible:

1. Service Health Check

# Check if service is running
curl -f http://localhost:8080/health || echo "Service not running"

# Check specific port
nc -z localhost 8080 && echo "Port open" || echo "Port closed"

# For Windows
Test-NetConnection -ComputerName localhost -Port 8080

2. Docker Compose Auto-Start (Preferred)

When an e2e/ directory exists with docker-compose files:

# Detect and start services in e2e directory
if [ -d "e2e" ]; then
  cd e2e && docker compose up -d
fi

# Or specific compose file
docker compose -f e2e/compose.yaml up -d

# Check service health status
docker compose -f e2e/compose.yaml ps

# View logs
docker compose -f e2e/compose.yaml logs -f

3. Project Structure Detection

Common e2e directory patterns:

project/
├── e2e/
│   ├── compose.yaml          # Main compose file
│   ├── compose-external.yaml # External services
│   ├── compose-k8s.yaml      # Kubernetes-specific
│   └── test-suite.yaml       # Test definitions
├── docker-compose.yaml       # Root level compose
└── Makefile                  # May contain test targets

4. Service Readiness Verification

# Wait for service health endpoint
while ! curl -f http://localhost:8080/health; do
  echo "Waiting for service..."
  sleep 2
done

# Check docker compose health status
docker compose -f e2e/compose.yaml ps --format json | \
  jq -r '.[] | select(.Health != "healthy" and .State == "running") | .Service'

# For services with health checks
docker compose -f e2e/compose.yaml ps | grep -q "healthy" || echo "Services not ready"

5. Environment Variables

# Check required environment variables
env | grep -E "TOKEN|PASSWORD|SECRET" || echo "Missing required env vars"

# Export required variables
export GITEE_TOKEN="your-token"
export GITHUB_TOKEN="your-token"

# Or use .env file
docker compose --env-file .env up -d

6. Cleanup After Tests

# Stop and remove services after testing
cd e2e && docker compose down

# Remove volumes (careful: deletes data)
docker compose -f e2e/compose.yaml down -v

# Keep services running for debugging
# docker compose -f e2e/compose.yaml ps

7. Pre-Test Workflow Example

# Complete pre-test workflow
#!/bin/bash
set -e

echo "🔍 Checking project structure..."
if [ -d "e2e" ]; then
  echo "📦 Found e2e directory, starting services..."
  cd e2e
  docker compose up -d

  echo "⏳ Waiting for services to be healthy..."
  timeout 60 bash -c 'until docker compose ps | grep -q "healthy"; do sleep 2; done'

  echo "✅ Services are ready"
  docker compose ps
  cd ..
else
  echo "⚠️  No e2e directory found, assuming service is already running"
fi

echo "🧪 Running tests..."
atest run -p testsuite.yaml

echo "🧹 Cleaning up..."
if [ -d "e2e" ]; then
  cd e2e && docker compose down
fi

Test Suite Properties

Suite Level

• name: Suite name
• api: Base API URL (supports templates: {{default "http://localhost:8080" (env "SERVER")}})
• param: Global parameters available in all tests
• spec: Protocol specification (for gRPC/GraphQL)
• items: Array of test cases

Test Case Level

Request:

• name: Test case name
• request.api: Endpoint path
• request.method: HTTP method (GET, POST, PUT, DELETE, etc.)
• request.header: Request headers
• request.body: Request body (supports templates)
• request.cookie: Cookies to send
• request.url: Full URL (overrides api + api combination)
• request.form: Form data
• request.files: File uploads

Expect:

• expect.statusCode: Expected HTTP status code
• expect.body: Expected response body
• expect.bodyFieldsExpect: Expected field values
• expect.schema: JSON schema for validation
• expect.verify: Array of verification expressions
• expect.verifyWithSelector: Verification with JSONPath
• expect.contentType: Expected content type

Control:

• before.items: Commands to run before test
• after.items: Commands to run after test
• cond: Conditional execution

gRPC Testing

name: grpc-sample
api: 127.0.0.1:7070
spec:
  kind: grpc
  rpc:
    import:
      - ./path/to/proto/files
    protofile: service.proto
items:
  - name: UnaryCall
    request:
      api: /service.Service/Method
      body: |
        {"field": "value"}
  - name: ServerStream
    request:
      api: /service.Service/StreamMethod
  - name: ClientStream
    request:
      api: /service.Service/Cl