How you can help

Most of what this work needs is not code. It needs people who read a language, use assistive technology, or can check a claim against the public record. Every task below is an open issue in a public repository.

If you read Hindi or another Indian language

Review by someone who reads the language is the capacity this workstream most needs, and it is something automated checks cannot supply.

• Hindi coverage gap in AppFlowy’s locale file (opens in a new tab): claim one section and translate or review it. AppFlowy accepts translation updates by pull request.
• AppFlowy’s hin.json locale-code mismatch (opens in a new tab): write it up as a short note the project could act on.
• Localization pipeline and case-study contributor (opens in a new tab): a standing role of about four to five hours a week.

If you use a screen reader or other assistive technology

These tasks need lived experience, not code. They are where the accessibility research is weakest without you.

• Validate the review rubric (opens in a new tab): are the criteria right, and what should count as assistive-technology evidence?
• What should a project’s review capacity promise? (opens in a new tab): the view of someone who uses assistive technology daily.
• The minimum meaningful assistive-technology test matrix (opens in a new tab): which combinations a reviewer should actually test.

If you like research: screening public pull requests

No prior accessibility or localization experience is needed for the tasks labelled tier-a.

• Find the next two or three localization case studies (opens in a new tab): screen public pull requests against the selection criteria.
• Find accessibility case studies 7 to 10 (opens in a new tab): the same method, for accessibility pull requests.
• Test one pattern against a non-Bootstrap project (opens in a new tab): does a finding from one project hold in another?
• Prior art on declaration files (opens in a new tab): how comparable files achieved, or failed to achieve, adoption.
• Corpus and evidence researcher (opens in a new tab): a standing role of about four to five hours a week.

If you follow AI contribution policies

Adding a catalogue entry is a single-file pull request of about fifteen minutes, from a template.

• Catalogue entry: Node.js (opens in a new tab): a ready-to-claim first entry.
• Map more existing AI-contribution policies (opens in a new tab): the catalogue has 21 entries and aims for 60.
• Is verification, not detection, the right primitive? (opens in a new tab): an open question for the draft schema.

If you maintain a project

• Run pip install i18n-security-lint or pip install readme-ci on your repository and report what they get wrong. False positives are bugs.
• Pilot the draft ai-contribution-policy.yml (opens in a new tab): in your own repository.
• Do these localization patterns match your review practice? (opens in a new tab): an open request for comment from people who review translations.

Not sure where to start?

Start here (opens in a new tab) for a short orientation, see the open volunteer listings (opens in a new tab), or tell us what you can help with and we will point you at something real.