Skip to content

Release: v0.30.0

Choose a tag to compare

@shinpr shinpr released this 06 Nov 04:19
· 4 commits to main since this release
d23f861

What's Changed

  • Large changes are reviewed in units of Work Plan tasks, with at most 3 code-reviewer calls running at once. Each reviewer reads its unit in full. Medium changes keep one reviewer.
  • code-reviewer no longer lists every acceptance criterion or maps the whole change in a fixed order. It judges each changed file and test against its review criteria, and still reports unfulfilled criteria as findings.
  • frontend-ai-guide is merged into ai-development-guide. The React/TypeScript-specific criteria now live in ai-development-guide/references/frontend.md, so shared agents such as code-reviewer and investigator also apply them.
  • Reviewers, investigators, and designers read the frontend and E2E references before they start. Executors and quality fixers read them at the step that uses them.
  • Task files list a test file only when the Work Plan fixes it, such as a generated skeleton. Executors choose the other tests by the behavior they prove.

Breaking Change

The frontend-ai-guide skill is removed. Use ai-development-guide and its references/frontend.md instead.