Implements continuous integration testing to ensure that workflows from the external githubnext/agentics repository can be successfully installed and compiled within the gh-aw system, with support for both automatic and manual workflow execution. Also includes build system improvements for better reproducibility.

Problem

The gh-aw project lacked automated testing to verify compatibility with external agentic workflow repositories. This created a risk of regressions where changes to gh-aw could break the ability to install and compile workflows from the broader agentics ecosystem, potentially going undetected until users encountered issues.

Additionally, TypeScript compilation errors in the JavaScript workflow files were preventing proper type checking and could lead to runtime issues. The build system was using npm install --package-lock-only which only updates the lock file without installing dependencies, making development setup inconsistent.

Solution

Integration Testing Workflow

Added a new GitHub Actions workflow at .github/workflows/integration-test-agentics.yml that:

• Triggers automatically on pushes to main branch and pull requests
• Supports manual triggering via workflow_dispatch for on-demand testing and debugging
• Tests end-to-end integration by installing the gh-aw CLI extension locally
• Validates external workflow compatibility by installing workflows from githubnext/agentics
• Ensures compilation works by running make recompile on all installed workflows
• Performs comprehensive validation using the mandatory make agent-finish command

TypeScript Compilation Fixes

Fixed TypeScript compilation issues in the JavaScript workflow files:

• Added proper JSDoc type annotations for parameter types in create_issue.cjs and create_pull_request.cjs
• Added type declarations for Node.js module imports (fs, crypto, child_process)
• Ensured CommonJS compatibility while providing proper TypeScript type checking

Build System Improvements

Updated the development dependency installation process:

• Changed make deps-dev to use npm ci instead of npm install --package-lock-only
• Provides reproducible builds by installing dependencies exactly as specified in package-lock.json
• Better suited for CI/CD environments and consistent development setups

Key Features

The workflow includes robust error handling and follows established CI patterns:

on:
  push:
    branches: [main]
  pull_request:
  workflow_dispatch:  # Enables manual triggering from GitHub UI

- name: Install workflows from agentics repository
  env:
    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
  run: |
    echo "Installing workflows from githubnext/agentics..."
    gh aw install githubnext/agentics
    echo "Successfully installed agentics workflows"

- name: Recompile all workflows
  run: |
    echo "Recompiling all installed workflows..."
    make recompile
    echo "Successfully recompiled all workflows"

The workflow leverages GitHub Actions' built-in authentication, eliminating the need for explicit GitHub CLI setup steps since runners are pre-authenticated with the provided GITHUB_TOKEN.

Benefits

• Early Detection: Catches integration issues before they reach users
• Ecosystem Compatibility: Ensures gh-aw remains compatible with external workflow repositories
• Regression Prevention: Automated testing prevents breaking changes to core compilation functionality
• Manual Testing Support: Allows developers to run integration tests on demand without creating commits
• Quality Assurance: Validates that the complete install → compile → validate pipeline works correctly
• Type Safety: Eliminates TypeScript compilation errors that could lead to runtime issues
• Reproducible Development: Consistent dependency installation across development environments
• Streamlined Authentication: Uses GitHub Actions' built-in authentication for cleaner CI execution

This addresses the need for comprehensive integration testing while maintaining the project's established CI patterns and quality standards, with additional improvements to the development workflow.

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.