Skip to content

Contributing

Arashi changes usually start in the arashi-arashi meta-repository, then land in the project repository that owns the code or docs. The goal is to keep work scoped, reviewable, and easy to coordinate across the Arashi workspace.

  1. Start from the meta-repo. Check workspace state with arashi status. If a repository is behind, pull the workspace forward before starting new work.
  2. Find or open the issue. Use the issue to agree on the problem, expected outcome, and affected repositories.
  3. Create a coordinated worktree. Run arashi create <branch-name> from arashi-arashi so every Arashi repository gets the same branch/worktree shape.
  4. Plan changes when the behavior or workflow shifts. For CLI behavior, cross-repo changes, or agent-facing workflows, use the OpenSpec proposal flow in arashi-arashi before implementation.
  5. Make implementation changes in the owning repository. For example, docs site content belongs in arashi-docs; CLI behavior belongs in arashi.
  6. Validate locally. Run the affected repository’s validation commands before opening a PR.
  7. Open focused PRs. Keep PRs small, link related PRs across repositories when needed, and reference the originating arashi-arashi issue.

If a change spans multiple repositories, open one PR per repository and cross-link them so reviewers can follow the full change.

Small documentation fixes can go directly through the docs repository. For larger docs restructuring, use the same contribution flow above and keep the page focused on user outcomes rather than internal site mechanics.

Docs maintainers may still need the site-maintenance references: