Skip to content

Block ambiguous branch and tag names

GitHub blocks branch and tag names which begin with refs/.

Under the hood, all Git refs begin with a prefix (refs/heads/ for branches and refs/tags/ for tags). In typical use, though, users rarely see these prefixes, so they're silently handled by GitHub, the Git client, and other tools. When a similar string is used as the beginning of the visible part of the branch or tag name, this results in ambiguity: did the user intend refs/heads/feature or refs/heads/refs/heads/feature? In nearly all cases, refs/ in front of a branch or tag name is accidental and becomes a problematic surprise later.

This change blocks new introductions of such names. Repositories with existing branches named this way can still push updates to those branches.

Code scanning API to enable default setup with CodeQL on a repository

Code scanning have shipped an API for repositories to programmatically enable code scanning default setup with CodeQL.

The API can be used to:

  • Onboard a repository to default setup: gh api -X PATCH /repos/[org-name]/[repo-name]/code-scanning/default-setup -f state=configured
  • Specify which CodeQL query suite to use in the default setup configuration: gh api -X PATCH /repos/[org-name]/[repo-name]/code-scanning/default-setup -f query_suite=extended
  • View the current default setup configuration for a repository: gh api /repos/[org]/[repo-name]/code-scanning/default-setup
  • Offboard a repository from default setup: gh api -X PATCH /repos/[org-name]/[repo-name]/code-scanning/default-setup -f state=not-configured

When you onboard a repository via the API, you will recieve a workflow run ID which can be used to monitor the setup progress. This can be used to see the status and conclusion of the run: gh api repos/[org-name]/[repo-name]/actions/runs/[run-id] --jq '.status, .conclusion'

{
  "state": "configured",
  "languages": ["javascript", "ruby"],
  "query_suite": "default", 
  "updated_at": "2023-02-24T20:00:42Z"
}

For more information, see "Get the code scanning default setup configuration" and "Update the code scanning default setup configuration".

See more

General availability of granular access token on npm

Today we are making the granular access token feature on npm generally available.

Granular access token, allows you to:

  • Restrict token access to specific packages and/or scopes
  • Grant tokens access to specific organizations for org and user management
  • Set a token expiration date
  • Limit token access based on IP address ranges
  • Select between read and/or write access for the token

We recommend using granular access tokens with least privileges for automating your publishing and org management activities. You can allow your package to be published without 2FA using granular access tokens from your trusted CI/CD systems. Additionally, you can also configure your package to require 2FA when publishing from a local machine to defend against account hijacking.

Read more about creating a granular access token here.

See more
View all changes