Skip to content
You're viewing an older version of this GitHub Action. Do you want to see the latest version instead?
shield

GitHub Action

Web Accessibility Evaluation

v2.1.8

Web Accessibility Evaluation

shield

Web Accessibility Evaluation

Fast and feature-packed web accessibility testing

Installation

Copy and paste the following snippet into your .yml file.

 
              
- name: Web Accessibility Evaluation

              
  uses: a11ywatch/[email protected]
Learn more about this action in a11ywatch/github-actions

Choose a version

github-action

Tests

A feature rich GitHub action that runs actionable accessibility reports on your website that can handle large workloads.

Some of the primary features include pass/fail testing, code fixes, and detailed reports.

When running locally the action uses A11yWatch Lite.

Usage

- uses: a11ywatch/[email protected]
  with:
    WEBSITE_URL: ${{ secrets.WEBSITE_URL }}
    SUBDOMAINS: true
    TLD: false
    SITEMAP: true
    FAIL_ERRORS_COUNT: 15
    LIST: true
    FIX: false
    UPGRADE: false
    UPLOAD: true
    COMPUTER_VISION_SUBSCRIPTION_KEY: ${{ secrets.COMPUTER_VISION_SUBSCRIPTION_KEY }}
    COMPUTER_VISION_ENDPOINT: ${{ secrets.COMPUTER_VISION_ENDPOINT }}
  env: 
    DEFAULT_RUNNERS: htmlcs,axe
    PAGEMIND_IGNORE_WARNINGS: true

Action inputs

All inputs are optional except $WEBSITE_URL.

Name Description Default
WEBSITE_URL Website domain to scan (Start with http:// or https://).
SITE_WIDE Site-wide scanning across all pages. false
FIX Attempt to apply recommendations to code and commit to github. false
SUBDOMAINS Include all subdomains (required SITE_WIDE=true). true
SITEMAP Extend crawl with sitemap links (required SITE_WIDE=true). true
TLD Include all tld extensions (required SITE_WIDE=true). true
LIST Report the results to github as a pass or fail list or detailed report. false
RECORD Record the audit as video to a directory.
FAIL_TOTAL_COUNT Determine whether to fail the CI if total issues warnings and errors exceed the counter. Takes precedence over the other FAIL inputs. 0
FAIL_ERRORS_COUNT Determine whether to fail the CI if total issues with errors exceed the counter. 0
FAIL_WARNINGS_COUNT Determine whether to fail the CI if total issues with warnings exceed the counter. 0
EXTERNAL Use the A11yWatch remote API for fast results. If this is set A11YWATCH_TOKEN is needed.
COMPUTER_VISION_SUBSCRIPTION_KEY Computer Vision API key for image recognition on alts.
COMPUTER_VISION_ENDPOINT Computer Vision url endpoint. false
DISABLE_PR_STATS Prevent messaging to the pr results of test. false
TOKEN GITHUB_TOKEN (permissions contents: write and pull-requests: write) or a repo scoped Personal Access Token (PAT). GITHUB_TOKEN
A11YWATCH_TOKEN The A11yWatch API token to use to identify a user.
SLIM Use the gRPC client to gather reports - only displays stats, useful for large websites (no code generation, no outputs, just pure stats) false
UPGRADE Upgrade the docker images before testing to latest. false
UPLOAD Upload the data as an artifact to get better insight. false

Action Outputs

Name Description Default
issues The amount of issues found

ENV Variables

Configure crawl wide settings for A11yWatch Lite (local) installs.

Name Description Default
DEFAULT_RUNNERS A comma separeted list of runners to use in Litemode for testing like axe, htmlcs, ace. htmlcs,axe
PAGEMIND_IGNORE_WARNINGS Enable to ignore all warnings from output. This could speed up runs and save on audit size. false

If you set it to only htmlcs you will have really fast crawls with good coverage with our fork.

Performance

On a larger website A11yWatch action runs over 60x-10,000x+ faster depending on CPUs/hardware. After the first installation you should have faster setup time between runs. You can expect to handle at least 1k pages per minute on a 2-core CPU 7 GB of RAM memory shared github action. If you enable RECORD the output time may increase a bit.

When AI_DISABLED is set to true the run for A11yWatch may increase.

Common Issues

If you experience issues on your CI you may have to toggle the UPGRADE input to true in order to get the latest installs. If you see a playwright error try adding PLAYWRIGHT_VERSION env variable with the newest version to install chrome.

CLI

In order to get the results of the run in json run results_json="$(a11ywatch -r)" across any step after the action.

LICENSE

check the license file in the root of the project.