gh-aw-test

End-to-end tests for GitHub Agentic Workflows (gh-aw). Tests AI-powered workflows that automatically create issues, pull requests, and comments.

Quick Start

1. Prerequisites: GitHub CLI, Git, authenticated GitHub account
2. Clone: git clone https://github.com/githubnext/gh-aw-test.git && cd gh-aw-test
3. Authenticate: gh auth login
4. Run tests: ./e2e.sh (gh-aw binary will be auto-built if needed)

Usage

Run all tests:

./e2e.sh

Run specific test types:

./e2e.sh --workflow-dispatch-only    # Direct workflow triggers
./e2e.sh --issue-triggered-only      # Tests via issue creation
./e2e.sh --command-triggered-only    # Tests via comments

Run specific tests:

./e2e.sh test-claude-*               # All Claude tests
./e2e.sh test-*-create-issue         # All issue creation tests
./e2e.sh --dry-run                   # Preview without running

Test Types

• Workflow dispatch: AI creates issues, PRs, code scanning alerts, uses MCP tools
• Issue-triggered: AI adds comments and labels, updates issues
• Command-triggered: AI responds to commands, pushes code, adds review comments

Automated Testing

Nightly GitHub Actions run at 3:00 AM UTC. Manually trigger via Actions → "Nightly E2E Tests". Logs are available as downloadable artifacts.

Troubleshooting

Authentication issues: Run gh auth login and gh auth status Missing gh-aw: Script auto-builds it, or manually: git clone https://github.com/githubnext/gh-aw.git && cd gh-aw && make build Timeouts: Check log file e2e-test-*.log for details