Skip to main content
The privacy policy says what CodeWarden reads, what it keeps, who else sees it, and how to have it deleted. If this page and the policy ever disagree, the policy is right; write to contact@codewarden.co so we can fix this page.

​
How the product is built

  • Nothing is cloned and nothing is executed. CodeWarden reads pull requests, files and .codewarden.yml through the GitHub API. It never runs a build, a test, a linter or any script from your repository. There is no setting that changes this.
  • Least GitHub access. The App asks for Contents: read, Pull requests: read and write, and Metadata: read. It does not ask for write access to your code.
  • No stored GitHub tokens. An installation token is requested for each job and never stored.
  • Read-only integrations. Jira is connected with read scopes only. Sentry uses a token you create with read permissions only. Both credentials are encrypted at rest and deleted when you disconnect.
  • Everything in a pull request is treated as untrusted. Titles, descriptions, code, comments and linked tickets reach the model as data, never as instructions. Rules come only from .codewarden.yml on the base branch, which a pull request cannot edit.
  • Every review is recorded. The prompt sent to the model, its answer, the comments posted and the ones dropped are kept so every comment can be explained. A Jira ticket’s text is sent to the model but not kept.

​
Your choices

For which services process your data, retention, backups, and the model provider’s data retention, read the privacy policy.