actions/_tests

GitHub Actions Testing Framework

A comprehensive testing framework for validating GitHub Actions in this monorepo using ShellSpec and Python-based input validation.

🚀 Quick Start

# Run all tests
make test

# Run only unit tests
make test-unit

# Run tests for specific action
make test-action ACTION=node-setup

# Run with coverage reporting
make test-coverage

Prerequisites

# Install ShellSpec (testing framework)
curl -fsSL https://github.com/shellspec/shellspec/releases/latest/download/shellspec-dist.tar.gz | tar -xz
sudo make -C shellspec-* install

# Install nektos/act (optional, for integration tests)
brew install act  # macOS
# or: curl https://raw.githubusercontent.com/nektos/act/master/install.sh | sudo bash

📁 Framework Overview

Architecture

The testing framework uses a multi-level testing strategy:

1. Unit Tests - Fast validation of action logic, inputs, and outputs using Python validation
2. Integration Tests - Test actions in realistic workflow environments
3. External Usage Tests - Validate actions work as ivuorinen/actions/action-name@main

Technology Stack

• Primary Framework: ShellSpec - BDD testing for shell scripts
• Validation: Python-based input validation via validate-inputs/validator.py
• Local Execution: nektos/act - Run GitHub Actions locally
• CI Integration: GitHub Actions workflows

Directory Structure

_tests/
├── README.md                    # This documentation
├── run-tests.sh                 # Main test runner script
├── unit/                       # Unit tests by action
│   ├── spec_helper.sh          # ShellSpec helper with validation functions
│   ├── version-file-parser/    # Example unit tests
│   ├── node-setup/            # Example unit tests
│   └── ...                    # One directory per action
├── framework/                   # Core testing utilities
│   ├── setup.sh                # Test environment setup
│   ├── utils.sh                # Common testing functions
│   ├── validation.py           # Python validation utilities
│   └── fixtures/               # Test fixtures
├── integration/               # Integration tests
│   ├── workflows/             # Test workflows for nektos/act
│   ├── external-usage/        # External reference tests
│   └── action-chains/         # Multi-action workflow tests
├── coverage/                  # Coverage reports
└── reports/                   # Test execution reports

✍️ Writing Tests

Basic Unit Test Structure

#!/usr/bin/env shellspec
# _tests/unit/my-action/validation.spec.sh

Describe "my-action validation"
ACTION_DIR="my-action"
ACTION_FILE="$ACTION_DIR/action.yml"

Context "when validating required inputs"
  It "accepts valid input"
    When call validate_input_python "my-action" "input-name" "valid-value"
    The status should be success
  End

  It "rejects invalid input"
    When call validate_input_python "my-action" "input-name" "invalid@value"
    The status should be failure
  End
End

Context "when validating boolean inputs"
  It "accepts true"
    When call validate_input_python "my-action" "dry-run" "true"
    The status should be success
  End

  It "accepts false"
    When call validate_input_python "my-action" "dry-run" "false"
    The status should be success
  End

  It "rejects invalid boolean"
    When call validate_input_python "my-action" "dry-run" "maybe"
    The status should be failure
  End
End
End

Integration Test Example

# _tests/integration/workflows/my-action-test.yml
name: Test my-action Integration
on: workflow_dispatch

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

      - name: Test action locally
        id: test-local
        uses: ./my-action
        with:
          required-input: 'test-value'

      - name: Validate outputs
        run: |
          echo "Output: ${{ steps.test-local.outputs.result }}"
          [[ -n "${{ steps.test-local.outputs.result }}" ]] || exit 1

      - name: Test external reference
        uses: ivuorinen/actions/my-action@main
        with:
          required-input: 'test-value'

🛠️ Testing Functions

Primary Validation Function

The framework provides one main validation function that uses the Python validation system:

validate_input_python

Tests input validation using the centralized Python validator:

validate_input_python "action-name" "input-name" "test-value"

Examples:

# Boolean validation
validate_input_python "pre-commit" "dry-run" "true"       # success
validate_input_python "pre-commit" "dry-run" "false"      # success
validate_input_python "pre-commit" "dry-run" "maybe"      # failure

# Version validation
validate_input_python "node-setup" "node-version" "18.0.0"       # success
validate_input_python "node-setup" "node-version" "v1.2.3"       # success
validate_input_python "node-setup" "node-version" "invalid"      # failure

# Token validation
validate_input_python "npm-publish" "npm-token" "ghp_123..."    # success
validate_input_python "npm-publish" "npm-token" "invalid"       # failure

# Docker validation
validate_input_python "docker-build" "image-name" "myapp"       # success
validate_input_python "docker-build" "tag" "v1.0.0"            # success

# Path validation (security)
validate_input_python "pre-commit" "config-file" "config.yml"   # success
validate_input_python "pre-commit" "config-file" "../etc/pass"  # failure

# Injection detection
validate_input_python "common-retry" "command" "echo test"      # success
validate_input_python "common-retry" "command" "rm -rf /; "     # failure

Helper Functions from spec_helper.sh

# Setup/cleanup
setup_default_inputs "action-name" "input-name"     # Set required defaults
cleanup_default_inputs "action-name" "input-name"   # Clean up defaults
shellspec_setup_test_env "test-name"               # Setup test environment
shellspec_cleanup_test_env "test-name"             # Cleanup test environment

# Mock execution
shellspec_mock_action_run "action-dir" key1 value1 key2 value2
shellspec_validate_action_output "expected-key" "expected-value"

# Action metadata
validate_action_yml "action.yml"          # Validate YAML structure
get_action_inputs "action.yml"           # Get action inputs
get_action_outputs "action.yml"          # Get action outputs
get_action_name "action.yml"             # Get action name

Complete Action Validation Example

Describe "comprehensive-action validation"
  ACTION_DIR="comprehensive-action"
  ACTION_FILE="$ACTION_DIR/action.yml"

  Context "when validating all input types"
    It "validates boolean inputs"
      When call validate_input_python "$ACTION_DIR" "verbose" "true"
      The status should be success

      When call validate_input_python "$ACTION_DIR" "verbose" "false"
      The status should be success

      When call validate_input_python "$ACTION_DIR" "verbose" "invalid"
      The status should be failure
    End

    It "validates numeric inputs"
      When call validate_input_python "$ACTION_DIR" "max-retries" "3"
      The status should be success

      When call validate_input_python "$ACTION_DIR" "max-retries" "999"
      The status should be failure
    End

    It "validates version inputs"
      When call validate_input_python "$ACTION_DIR" "tool-version" "1.0.0"
      The status should be success

      When call validate_input_python "$ACTION_DIR" "tool-version" "v1.2.3-rc.1"
      The status should be success
    End

    It "validates security patterns"
      When call validate_input_python "$ACTION_DIR" "command" "echo test"
      The status should be success

      When call validate_input_python "$ACTION_DIR" "command" "rm -rf /; "
      The status should be failure
    End
  End

  Context "when validating action structure"
    It "has valid YAML structure"
      When call validate_action_yml "$ACTION_FILE"
      The status should be success
    End
  End
End

🎯 Testing Patterns by Action Type

Setup Actions (node-setup, php-version-detect, etc.)

Focus on version detection and environment setup:

Context "when detecting versions"
  It "detects version from config files"
    When call validate_input_python "node-setup" "node-version" "18.0.0"
    The status should be success
  End

  It "accepts default version"
    When call validate_input_python "python-version-detect" "default-version" "3.11"
    The status should be success
  End
End

Linting Actions (eslint-fix, prettier-fix, etc.)

Focus on file processing and security:

Context "when processing files"
  It "validates working directory"
    When call validate_input_python "eslint-fix" "working-directory" "."
    The status should be success
  End

  It "rejects path traversal"
    When call validate_input_python "eslint-fix" "working-directory" "../etc"
    The status should be failure
  End

  It "validates boolean flags"
    When call validate_input_python "eslint-fix" "fix-only" "true"
    The status should be success
  End
End

Build Actions (docker-build, go-build, etc.)

Focus on build configuration:

Context "when building"
  It "validates image name"
    When call validate_input_python "docker-build" "image-name" "myapp"
    The status should be success
  End

  It "validates tag format"
    When call validate_input_python "docker-build" "tag" "v1.0.0"
    The status should be success
  End

  It "validates platforms"
    When call validate_input_python "docker-build" "platforms" "linux/amd64,linux/arm64"
    The status should be success
  End
End

Publishing Actions (npm-publish, docker-publish, etc.)

Focus on credentials and registry validation:

Context "when publishing"
  It "validates token format"
    When call validate_input_python "npm-publish" "npm-token" "ghp_123456789012345678901234567890123456"
    The status should be success
  End

  It "rejects invalid token"
    When call validate_input_python "npm-publish" "npm-token" "invalid-token"
    The status should be failure
  End

  It "validates version"
    When call validate_input_python "npm-publish" "package-version" "1.0.0"
    The status should be success
  End
End

🔧 Running Tests

Command Line Interface

# Basic usage
./_tests/run-tests.sh [OPTIONS] [ACTION_NAME...]

# Examples
./_tests/run-tests.sh                           # All tests, all actions
./_tests/run-tests.sh -t unit                   # Unit tests only
./_tests/run-tests.sh -a node-setup             # Specific action
./_tests/run-tests.sh -t integration docker-build  # Integration tests for docker-build
./_tests/run-tests.sh --format json --coverage  # JSON output with coverage

Options

Option	Description
-t, --type TYPE	Test type: unit, integration, e2e, all
-a, --action ACTION	Filter by action name pattern
-j, --jobs JOBS	Number of parallel jobs (default: 4)
-c, --coverage	Enable coverage reporting
-f, --format FORMAT	Output format: console, json, junit
-v, --verbose	Enable verbose output
-h, --help	Show help message

Make Targets

make test                    # Run all tests
make test-unit              # Unit tests only
make test-integration       # Integration tests only
make test-coverage          # Tests with coverage
make test-action ACTION=name # Test specific action

🤝 Contributing Tests

Adding Tests for New Actions

1. Create Unit Test Directory

   mkdir -p _tests/unit/new-action
   

2. Write Unit Tests

   # _tests/unit/new-action/validation.spec.sh
   #!/usr/bin/env shellspec
   
   Describe "new-action validation"
   ACTION_DIR="new-action"
   ACTION_FILE="$ACTION_DIR/action.yml"
   
   Context "when validating inputs"
     It "validates required input"
       When call validate_input_python "new-action" "required-input" "value"
       The status should be success
     End
   End
   End
   

3. Create Integration Test

   # _tests/integration/workflows/new-action-test.yml
   # (See integration test example above)
   

4. Test Your Tests

   make test-action ACTION=new-action
   

Pull Request Checklist

• Tests use validate_input_python for input validation
• All test types pass locally (make test)
• Integration test workflow created
• Security testing included for user inputs
• Tests are independent and isolated
• Proper cleanup in test teardown
• Documentation updated if needed

💡 Best Practices

1. Use validate_input_python for All Input Testing

✅ Good:

When call validate_input_python "my-action" "verbose" "true"
The status should be success

❌ Avoid:

# Don't manually test validation - use the Python validator
export INPUT_VERBOSE="true"
python3 validate-inputs/validator.py

2. Group Related Validations

✅ Good:

Context "when validating configuration"
  It "accepts valid boolean"
    When call validate_input_python "my-action" "dry-run" "true"
    The status should be success
  End

  It "accepts valid version"
    When call validate_input_python "my-action" "tool-version" "1.0.0"
    The status should be success
  End
End

3. Always Include Security Testing

✅ Always include:

It "rejects command injection"
  When call validate_input_python "common-retry" "command" "rm -rf /; "
  The status should be failure
End

It "rejects path traversal"
  When call validate_input_python "pre-commit" "config-file" "../etc/passwd"
  The status should be failure
End

4. Write Descriptive Test Names

✅ Good:

It "accepts valid semantic version format"
It "rejects version with invalid characters"
It "falls back to default when no version file exists"

❌ Avoid:

It "validates input"
It "works correctly"

5. Keep Tests Independent

• Each test should work in isolation
• Don't rely on test execution order
• Clean up after each test
• Use proper setup/teardown

🔍 Framework Features

Test Environment Setup

The framework automatically sets up test environments via spec_helper.sh:

# Automatic setup on load
- GitHub Actions environment variables
- Temporary directories
- Mock GITHUB_OUTPUT files
- Default required inputs for actions

# Available variables
$PROJECT_ROOT        # Repository root
$TEST_ROOT          # _tests/ directory
$FRAMEWORK_DIR      # _tests/framework/
$FIXTURES_DIR       # _tests/framework/fixtures/
$TEMP_DIR           # Temporary test directory
$GITHUB_OUTPUT      # Mock outputs file
$GITHUB_ENV         # Mock environment file

Python Validation Integration

All input validation uses the centralized Python validation system from validate-inputs/:

• Convention-based automatic validation
• 9 specialized validators (Boolean, Version, Token, Numeric, File, Network, Docker, Security, CodeQL)
• Custom validator support per action
• Injection and security pattern detection

🚨 Troubleshooting

Common Issues

"ShellSpec command not found"

# Install ShellSpec globally
curl -fsSL https://github.com/shellspec/shellspec/releases/latest/download/shellspec-dist.tar.gz | tar -xz
sudo make -C shellspec-* install

"act command not found"

# Install nektos/act (macOS)
brew install act

# Install nektos/act (Linux)
curl https://raw.githubusercontent.com/nektos/act/master/install.sh | sudo bash

Tests timeout

# Increase timeout for slow operations
export TEST_TIMEOUT=300

Permission denied on test scripts

# Make test scripts executable
find _tests/ -name "*.sh" -exec chmod +x {} \;

Debugging Tests

1. Enable Verbose Mode

   ./_tests/run-tests.sh -v
   

2. Run Single Test

   shellspec _tests/unit/my-action/validation.spec.sh
   

3. Enable Debug Mode

   export SHELLSPEC_DEBUG=1
   shellspec _tests/unit/my-action/validation.spec.sh
   

4. Check Test Output

   # Test results stored in _tests/reports/
   cat _tests/reports/unit/my-action.txt
   

📚 Resources

• ShellSpec Documentation
• nektos/act Documentation
• GitHub Actions Documentation
• Testing GitHub Actions Best Practices
• validate-inputs Documentation

Framework Development

Framework File Structure

_tests/
├── unit/
│   └── spec_helper.sh          # ShellSpec configuration and helpers
├── framework/
│   ├── setup.sh                # Test environment initialization
│   ├── utils.sh                # Common utility functions
│   ├── validation.py           # Python validation helpers
│   └── fixtures/               # Test fixtures
└── integration/
    ├── workflows/              # Integration test workflows
    ├── external-usage/         # External reference tests
    └── action-chains/          # Multi-action tests

Available Functions

From spec_helper.sh (_tests/unit/spec_helper.sh):

• validate_input_python(action, input_name, value) - Main validation function
• setup_default_inputs(action, input_name) - Set default required inputs
• cleanup_default_inputs(action, input_name) - Clean up default inputs
• shellspec_setup_test_env(name) - Setup test environment
• shellspec_cleanup_test_env(name) - Cleanup test environment
• shellspec_mock_action_run(action_dir, ...) - Mock action execution
• shellspec_validate_action_output(key, value) - Validate outputs

From utils.sh (_tests/framework/utils.sh):

• validate_action_yml(file) - Validate action YAML
• get_action_inputs(file) - Extract action inputs
• get_action_outputs(file) - Extract action outputs
• get_action_name(file) - Get action name
• test_input_validation(dir, name, value, expected) - Test input
• test_action_outputs(dir) - Test action outputs
• test_external_usage(dir) - Test external usage

Last Updated: October 15, 2025