A pull request limit can appear to solve a repository moderation problem until contributors start opening draft pull requests instead of regular ones. Maintainers configure a maximum number of open pull requests per contributor, yet the review queue continues to fill with unfinished changes, duplicate submissions, and low-quality contributions. The reason is that draft pull requests were previously excluded from GitHub's per-user pull request limit.

GitHub addressed this gap on October 8, 2026, by introducing a setting that allows repository maintainers to include draft pull requests when enforcing these limits. The change gives maintainers more control over how many concurrent pull requests contributors without write access can open, including submissions that are not yet marked ready for review.

The distinction matters for open-source projects that receive contributions from many external developers. Draft pull requests are useful for sharing work in progress, but they can also be used to bypass a limit intended to control contribution volume. With the new setting, repositories can decide whether unfinished submissions should consume the same allowance as regular open pull requests.

What Changed in GitHub Pull Request Limits?

GitHub introduced pull request limits in June 2026 to help maintainers control the number of concurrent open pull requests submitted by users without write access. When a contributor reaches the configured limit, they must close or merge an existing pull request before opening another one.

Initially, draft pull requests did not count toward that limit. A contributor could therefore open multiple drafts without consuming the allowance configured for ordinary pull requests. That behavior was useful for developers who wanted to share early work without occupying the main review queue, but it also created a loophole for repositories dealing with spam or repeated low-quality submissions.

The October update makes draft handling configurable. Repository maintainers can enable the option to count drafts toward a user's pull request limit rather than relying exclusively on the previous behavior.

For example, suppose a public repository allows each external contributor to have three open pull requests. Under the previous behavior, a contributor could have three regular open pull requests and additional drafts because those drafts were excluded from the limit. With draft counting enabled, every qualifying open pull request, including drafts, consumes the configured allowance.

This is a change to the repository's contribution policy, not a new limit on the total number of pull requests GitHub can store.

Why Draft Pull Requests Can Become a Moderation Problem

Draft pull requests serve a legitimate purpose. A developer can share incomplete implementation work, ask for early architectural feedback, or open a contribution before all tests and documentation are ready. Reviewers can provide guidance without treating the change as ready to merge.

The problem appears when the number of open submissions grows faster than maintainers can review them. An external contributor might open separate drafts for several experimental approaches, leave outdated attempts open, or repeatedly submit unfinished changes without addressing previous feedback. Automated tools can also create large numbers of submissions if their workflows are poorly controlled.

Even when these pull requests are not ready for formal review, they can create operational overhead. Maintainers must distinguish active work from abandoned attempts, track duplicate changes, and determine which submissions deserve attention. Depending on repository automation, opening additional pull requests can also trigger notifications, workflow executions, or other processing.

A configured pull request limit helps control this volume, but it only works as intended if the types of submissions that matter to the repository are included in the count.

The new draft-counting option lets maintainers choose between two policies. Projects that encourage early collaboration can continue excluding drafts. Projects that need tighter control over concurrent submissions can count drafts against the same allowance as regular open pull requests.

How to Enable Draft Pull Request Counting

Repository administrators can configure the behavior through GitHub's repository settings.

  1. Open the repository on GitHub.

  2. Navigate to Settings.

  3. Under the moderation options, select Interaction limits.

  4. Locate the Pull request limits section.

  5. Configure the maximum number of concurrent open pull requests allowed for users without write access.

  6. Enable the option to count draft pull requests toward the limit.

  7. Save or apply the configuration if GitHub presents a confirmation action.

The setting is part of the pull request creation limit. Enabling draft counting without configuring an applicable limit does not establish a meaningful maximum by itself.

The limit applies to users without write access in the relevant public repository. Contributors with write access or higher are not restricted by this limit, and trusted contributors can be placed on the bypass list when they need an exception.

Organizations can also configure pull request limits across their public repositories. Teams managing multiple repositories should review their organization-level policy as well as repository-specific settings, because the applicable scope determines which contribution rules take effect.

Configure Draft Counting Through the REST API

Repositories that manage moderation settings programmatically can also configure pull request creation limits through GitHub's REST API. The pull request creation-cap endpoint accepts an include_drafts Boolean field.

The following example shows the relevant request structure:

Bash
PATCH /repos/OWNER/REPO/interaction-limits/pulls/creation-cap
Accept: application/vnd.github+json
Authorization: Bearer YOUR_TOKEN
X-GitHub-Api-Version: 2026-03-10
Content-Type: application/json

{
  "enabled": true,
  "max_open_pull_requests": 3,
  "include_drafts": true
}

Replace OWNER and REPO with the repository owner and repository name. Use an appropriately authorized token, and store credentials securely rather than committing them to source control.

The three fields control different aspects of the policy:

  • enabled determines whether the pull request creation limit is active.

  • max_open_pull_requests sets the maximum number of concurrent open pull requests allowed per affected user.

  • include_drafts determines whether draft pull requests count toward that limit.

In this example, the limit is enabled, the maximum is three, and drafts are included in the count. This illustrates the request configuration; the appropriate threshold depends on the repository's contribution model.

The API is useful when teams want repeatable repository configuration or need to apply similar policies across multiple repositories. However, automation should verify the response and report configuration failures rather than assuming a request succeeded.

Do not confuse this endpoint with a GitHub API rate limit or a limit on the total number of pull requests in a repository. It controls how many qualifying open pull requests an affected contributor may have at one time.

What Happens When a Contributor Reaches the Limit?

Consider a public open-source repository with a limit of three concurrent open pull requests. A contributor has two regular pull requests open and creates a third draft.

With draft counting enabled, the contributor has reached the configured maximum. Opening another pull request requires closing or merging an existing one first. The contributor cannot avoid the restriction merely by marking new work as a draft.

The same rule applies if the contributor has three drafts open and then attempts to open a regular pull request. All qualifying open submissions consume the allowance, so the contributor must free a slot before creating another pull request.

This behavior encourages contributors to maintain a manageable set of active submissions. It also gives maintainers a more consistent way to control contribution volume without requiring them to review every unfinished change individually.

However, the policy has limits. It does not establish a quality threshold for submissions, prevent every form of spam, or guarantee that CI usage will decrease. Its effectiveness depends on the type of unwanted activity a repository is experiencing and how much of that activity comes from users subject to the configured limit.

Important Considerations for Repository Maintainers

Choose a limit that supports the contribution workflow

A low limit can reduce noise, but it can also prevent a legitimate contributor from working on several independent issues. A high limit may preserve flexibility but provide less protection against repeated submissions.

Review the repository's normal contribution patterns before choosing a maximum. Projects that encourage small, independent changes may need a different threshold from projects where contributors routinely work across several packages or branches.

Decide whether drafts should count

Counting drafts is particularly useful when a repository receives repeated unfinished submissions or when draft creation is being used to bypass contribution limits.

On the other hand, some projects explicitly encourage contributors to open early drafts for architectural discussion. Excluding drafts may make more sense when early collaboration is central to the project's review process.

The setting makes this a repository policy decision rather than forcing every project into the same workflow.

Use bypasses deliberately

GitHub supports a bypass list for trusted contributors who need to exceed a configured limit without receiving additional repository permissions. This can be useful for frequent external contributors who regularly maintain multiple open changes.

A bypass should reflect a genuine workflow requirement, not serve as a substitute for choosing an appropriate limit. Keep the list focused on contributors who need the exception and review it as project responsibilities change.

Test API-managed settings

Teams that configure interaction limits through scripts or administrative automation should check the HTTP response and confirm the resulting setting. Invalid values, insufficient permissions, or an incorrect endpoint can cause an update to fail.

Also keep the configuration in a controlled administrative workflow. A moderation policy should not be silently altered by an unrelated deployment script or an automation process that has broader permissions than necessary.

Common Mistakes to Avoid

Assuming every draft pull request is automatically counted. The behavior is configurable. Maintainers must enable draft counting for the applicable limit rather than assuming the October update automatically changes every repository's policy.

Treating the limit as a total repository quota. The setting restricts concurrent open pull requests for users without write access. It does not cap the total number of pull requests stored in a repository.

Expecting the limit to reduce all CI activity. Drafts may still be legitimate and workflows can run for other reasons. The setting can reduce unnecessary processing associated with excessive submissions, but the exact effect depends on the repository's automation and contribution patterns.

Ignoring trusted contributors' requirements. A strict limit can interfere with legitimate work when contributors need to maintain multiple independent changes. Review the bypass list and the configured threshold instead of increasing permissions unnecessarily.

Confusing draft status with closure. A draft is still an open pull request. When draft counting is enabled, converting a regular pull request to draft status does not remove it from the count. Contributors who reach the limit must close or merge an existing pull request to free a slot.

Advantages and Disadvantages

Advantages

  • More effective contribution controls: Counting drafts closes the previous gap that allowed contributors to open additional draft pull requests beyond their configured allowance.

  • Less review-queue noise: Repositories can limit the number of concurrent submissions from users without write access, including work that is not ready for formal review.

  • Flexible moderation policies: Maintainers can decide whether drafts should count, preserving early collaboration where appropriate while enabling stricter limits elsewhere.

  • Better administrative consistency: Teams can configure the policy through repository settings or the REST API, making it easier to manage moderation rules deliberately.

Disadvantages

  • Less flexibility for experimental work: Contributors who frequently share multiple early drafts may need to close or merge existing pull requests before opening additional ones.

  • Limited effect on trusted or internal contributors: The limit does not apply to users with write access or higher, and authorized bypasses can exempt selected contributors.

  • Not a complete spam solution: The feature controls concurrent submissions but does not determine whether a pull request contains useful code or prevent every unwanted contribution.

Summary

GitHub's October 8, 2026, update allows maintainers to configure pull request limits so that draft pull requests count toward a contributor's maximum number of concurrent open submissions. Previously, drafts were excluded, creating a gap that could be exploited to open additional submissions even after reaching the configured limit.

For public repositories that receive large volumes of external contributions, enabling draft counting can help reduce review-queue clutter and constrain unnecessary notifications or automation triggered by excessive pull requests. Projects that depend on early draft reviews can retain the previous behavior by leaving the option disabled.

The practical decision is to align the limit with the repository's contribution workflow. Choose a reasonable threshold, decide whether unfinished submissions should consume an allowance, and use bypasses only for contributors who genuinely require them. The setting is a focused moderation control, not a replacement for good contribution guidelines, review practices, or CI workflow design.