CI & reports

Yellow Jacket uses standard process exit codes and works with normal CI systems.

The same route inventory and coverage rules used locally can be reused in automation.

Coverage gate

Run:

yellow-jacket coverage

With a configured minimum:

coverage: {
  openapi:
    './openapi.json',

  minimum:
    80
}

the command returns:

exit 0 = coverage requirement satisfied
exit 1 = coverage requirement not satisfied

JSON output

For machines and further tooling:

yellow-jacket coverage \
  --json \
  --output yellow-jacket-coverage.json

Markdown output

For human-readable CI summaries:

yellow-jacket coverage \
  --markdown \
  --output yellow-jacket-coverage.md

GitHub Actions

A complete example is included in:

examples/ci/github-actions.yml

The report can be appended to:

$GITHUB_STEP_SUMMARY

while preserving Yellow Jacket's original exit code.

GitLab CI

A GitLab example is included in:

examples/ci/gitlab-ci.yml

The Markdown report can be preserved as a job artifact even when the coverage gate fails.

HTTP regression reporting

HTTP runs support the same generic report formats as coverage.

JSON:

yellow-jacket run \
  --json \
  --output yellow-jacket-run.json

Markdown:

yellow-jacket run \
  --markdown \
  --output yellow-jacket-run.md

Static HTML:

yellow-jacket run \
  --html \
  --output yellow-jacket-run.html

The JSON report contains the run result, individual request results and detected regressions.

Configured compare.ignore and compare.redact rules are applied to response bodies before JSON output is serialized, so values configured for redaction are not exposed by the report.

All formats preserve the normal run exit code.

The native CI formats also work with the main regression runner.

GitHub Actions:

yellow-jacket run --github

Failed expectations and detected baseline regressions are emitted as native workflow errors.

This includes regressions in:

• status codes
• response bodies
• selected response headers
• redirect behavior
• final redirect URLs

GitLab:

yellow-jacket run \
  --gitlab \
  --output yellow-jacket-run.xml

The resulting file is a JUnit report and can be published with GitLab artifacts:reports:junit.

If no baseline exists, request expectations are still checked and the native report explicitly states that regression comparison was skipped.

Native GitHub Actions reporting

Use:

yellow-jacket coverage --github

Yellow Jacket emits GitHub Actions workflow annotations.

Uncovered operations are reported as warnings. If the configured coverage minimum is not satisfied, the coverage summary is emitted as an error and the command exits with code 1.

When GITHUB_STEP_SUMMARY is available, Yellow Jacket also appends its Markdown coverage report to the job summary automatically.

This means a GitHub Actions step can be as small as:

- name: Yellow Jacket coverage
  run: npx yellow-jacket coverage --github

Native GitLab reporting

Use the GitLab format to generate a JUnit report:

yellow-jacket coverage \
  --gitlab \
  --output yellow-jacket-junit.xml

Then expose it as a GitLab test report:

artifacts:
  when: always

  reports:
    junit:
      - yellow-jacket-junit.xml

Covered operations are represented as passing test cases. Uncovered operations are represented as skipped test cases. The synthetic coverage gate fails only when the configured minimum is not satisfied.

Static HTML reports

Generate an HTML report:

yellow-jacket coverage \
  --html \
  --output public/index.html

The resulting file is fully static.

It can be:

• opened locally
• uploaded as a CI artifact
• served by a static hosting service
• published with GitHub Pages

GitHub Pages

A consumer-project example is available in:

examples/ci/github-pages.yml

It generates:

public/index.html

uploads it as a GitHub Pages artifact and deploys it.

The workflow preserves the original coverage status so a failed coverage gate can still publish a report for inspection.

Yellow Jacket documentation

The Yellow Jacket project documentation itself is also built as a static site with VitePress and deployed using GitHub Pages.