CI and deployment pipelines

Create and review a suite in AegisRunner before using it as a build check. CI triggers require a Pro or Business plan and a token for the project's tests.

Prepare the token and suite

1. Open the project, then Project setup → CI/CD.
2. Generate a named CI token and copy it when shown.
3. Store it in your CI secret store as AEGIS_TOKEN.
4. Choose the suite ID you want the pipeline to run.

Use separate tokens for independent pipelines so that you can revoke one without disrupting the others.

GitHub Actions example

Save this as .github/workflows/aegisrunner.yml in the application repository. Configure AEGIS_TOKEN as a repository secret and AEGIS_SUITE_ID as a repository variable.

name: AegisRunner regression
on:
  workflow_dispatch:
  push:
    branches: [main]
permissions:
  contents: read
jobs:
  test:
    runs-on: ubuntu-latest
    timeout-minutes: 35
    steps:
      - uses: actions/setup-node@v4
        with:
          node-version: '22'
      - name: Install AegisRunner CLI
        run: npm install -g @aegisrunner/cli@1.0.28
      - name: Run suite
        env:
          AEGIS_TOKEN: ${{ secrets.AEGIS_TOKEN }}
          AEGIS_SUITE_ID: ${{ vars.AEGIS_SUITE_ID }}
        run: |
          test -n "$AEGIS_SUITE_ID" || { echo "Set AEGIS_SUITE_ID"; exit 2; }
          aegis run --suite "$AEGIS_SUITE_ID" --timeout 1800 --format junit --output aegis-results/junit.xml
      - name: Preserve results
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: aegisrunner-results
          path: aegis-results/junit.xml
          if-no-files-found: warn

The target deployment must already be reachable when this job runs. Place the test job after your deployment job if both live in one workflow.

The CLI returns a nonzero exit code for failures, errors, and cancellations. Uploading with always() preserves a report when one was produced even if the test step failed.

Other CI systems

Install the same CLI in GitLab CI, Jenkins, or another Node-capable runner. Inject the token through its secret store, run aegis run, and publish the JUnit file.

Do not expose a project token to untrusted pull request code.

Scans versus regression checks

Use aegis scan --watch to discover app changes and generate tests. Use aegis run to execute existing tests. Keep these as explicit pipeline steps so the expected completion and failure behavior is clear.

For HTTP integration without the CLI, see API overview.