Uso seguro de pull_request_target

This guide helps you assess whether your workflow should use the pull_request_target event and understand the security risks involved. It also explains the protections that GitHub applies to these risks by default, and how to opt out of those protections if necessary.

Risks of pull_request_target

Workflows triggered by pull_request_target run with elevated trust: the job receives the base repository's GITHUB_TOKEN and access to repository and organization secrets. This is the same trust given to events like push that only collaborators can trigger, and it is what makes pull_request_target useful for automation that responds to pull requests from forks, such as labeling, triage, or for posting authenticated status checks.

To understand why this is safe by default, and how that safety is commonly broken, review pull_request_target against pull_request.

The pull_request event (along with pull_request_review and pull_request_review_comment) is unusual: it runs the workflow file from the merge commit of the pull request. For a pull request opened from a fork, that commit is controlled by someone without write access to the base repository. To run untrusted workflow code safely, GitHub restricts these events to a read-only GITHUB_TOKEN, withholds access to other secrets, and applies fork approval policies to prevent compute abuse. By default, actions/checkout in a pull_request workflow also checks out the pull request's merge commit, so the code checked out and the workflow that runs are consistent.

pull_request_target makes one critical and subtle change: the workflow, and any subsequent actions/checkout call that does not specify a ref, is taken from the base repository's default branch, not from the pull request. Because only trusted code from the default branch runs, it is safe to grant secrets and a read/write token. No code from the fork is executed by default.

You introduce risk when a workflow author overrides this default to run the fork's code. Developers frequently choose pull_request_target because they want to run a fork's pull request through CI and have access to secrets, for example to run tests that need a private registry. To do this, they point actions/checkout at the pull request head instead of the default branch, which is insecure:

# INSECURE. Provided as an example only.
on:
  pull_request_target:

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v6
        with:
          ref: ${{ github.event.pull_request.head.sha }}
      - name: Test
        run: make test

The checkout step alone does not execute untrusted code. The workflow file itself still comes from the default branch. The vulnerability is completed by the next step that runs code checked out into the current working directory. Here, make test executes a Makefile taken from the pull request head. An attacker only needs to open a pull request from a fork whose Makefile (or build script, test command, dependency, or configuration file) contains malicious commands. Those commands then run with the base repository's secrets and token.

This pattern is known as a "pwn request" and has been the root cause of multiple supply-chain compromises. For more information, see Preventing pwn requests from the GitHub Security Lab. Common vulnerable shapes include:

• Checking out a pull request's head or merge commit in actions/checkout (ref: ${{ github.event.pull_request.head.sha }}, ref: refs/pull/${{ github.event.pull_request.number }}/merge) and then building, testing, or otherwise executing the result.
• Setting repository: to the fork (repository: ${{ github.event.pull_request.head.repo.full_name }}) to pull the fork's branch directly.
• Fetching the pull request code outside of actions/checkout (for example with git fetch, gh pr checkout, or by downloading an artifact from a fork's pull_request run) and then running it.

Pwn requests are also not unique to pull_request_target. Any event that runs with secrets can introduce a pwn request if it checks out or downloads and executes untrusted code. For example, an issue_comment or workflow_run workflow that fetches and runs a fork's pull request code is vulnerable in the same way. A workflow_run workflow should treat artifacts uploaded by other workflows as untrusted data, since their contents can come from a fork.

Default policy for pull_request_target

To help protect your workflows from untrusted pull requests, GitHub provides a default event policy that blocks the pull_request_target event in public repositories.

How the default policy works

For public repositories that do not already have an applicable Actions event policy, GitHub adds a default policy that blocks workflows triggered by pull_request_target. For more information about policies, see About Actions policies.

The default policy:

• Does not apply to private or internal repositories.
• Does not replace an applicable event policy that you have already configured.
• Currently runs in evaluate mode. In this mode, workflow runs continue, but you can use policy insights to identify runs that would be blocked after enforcement.

On November 2, 2026, GitHub will enforce the default policy for affected repositories that were using the default pull_request_target policy before general availability.

Reviewing the impact before enforcement

While the policy is in evaluate mode, review your policy insights to identify workflows that use pull_request_target and would be blocked when the policy is enforced.

For each affected workflow, decide whether pull_request_target is still required:

• If the workflow does not need pull_request_target, update it to use a safer event where appropriate, such as pull_request.
• If the workflow must continue to use pull_request_target, create or update an applicable Actions event policy that explicitly allows pull_request_target.
• If you do not want to allow pull_request_target, leave the default policy in place. After enforcement, GitHub will block workflows triggered by that event.

For more information about configuring event policies and viewing insights, see Controlling who can execute GitHub Actions workflows. To manage policies programmatically, see REST API endpoints for GitHub Actions policies.

Deciding whether to use pull_request_target

Some workflows need to check out fork pull request code with elevated trust, and this is why pull_request_target was created in the first place. For example, generating coverage reports that require a private artifact registry or producing and running authenticated checks against the changes introduced from the pull request. Consider the questions below before using pull_request_target or opting into the allow-unsafe-pr-checkout flag in actions/checkout.

• Can you use pull_request instead? pull_request triggers on the same events as pull_request_target and runs the workflow code from the pull_request merge branch. It does this safely on pull requests from forks with the protections detailed above. If additional secret access is not needed, use pull_request. More complex workflows can be restructured to separate potentially dangerous handling of pull request code from accessing secrets. For more information, see Preventing pwn requests from the GitHub Security Lab.

• Is the checked-out code ever executed? This is the flaw that introduces pwn request vulnerabilities. It is most commonly introduced with actions/checkout by checking out a pull request head into the working directory and then running it. Unless the path input is set, actions/checkout writes the code into the $GITHUB_WORKSPACE directory, which is typically the working directory where subsequent commands run. Execution is not limited to your own steps: build and test commands such as npm install and npm run build, as well as configuration files and dependencies the code brings with it, can all run attacker-controlled code. Execution does not require an obvious build step. You must ensure the checked-out code is only ever inspected as data and never executed before using a pull_request_target event.

Hardening a pull_request_target workflow

If you have confirmed you need pull_request_target, apply these controls to limit the impact of this high-risk event. These apply whether or not your workflow checks out pull request code.

• Restrict secrets. Confirm that the permissions set on the GITHUB_TOKEN have the least privileges and that only the necessary repository and organization secrets are used for the workflow. For more information, see Use GITHUB_TOKEN for authentication in workflows.

• Understand the impact to caching. To reduce the risk of cache poisoning, workflows triggered by pull_request_target have read-only access to the cache in the default branch's scope. These workflows can restore existing cache entries but cannot create or overwrite them, so they cannot affect the execution of other, unrelated, workflows through the shared cache. If such a workflow attempts to save a cache, the save fails but the step and the job continue, and the failure is reported as a warning in the workflow log. If your workflow needs to populate the cache, save it from a workflow that runs on a trusted trigger such as push. A workflow or job can opt out of this read-only restriction by explicitly declaring a write-capable cache-mode, but doing so on a pull_request_target workflow reintroduces the cache-poisoning risk this restriction is designed to prevent. For more information, see Dependency caching reference.

• Ensure the underlying compute is isolated and ephemeral. If self-hosted runners are used, you must confirm that the runner environment is properly restricted from internal resources and is not reused across GitHub Actions runs. For more information, see Secure use reference.

• Enforce GitHub Actions security best practices. In addition to the specific risks of pwn requests, other common vulnerabilities, such as command injection, can exist and impact the code executed in this privileged event. For more information, see Keeping your GitHub Actions and workflows secure: Untrusted input from the GitHub Security Lab. To identify and proactively protect against common GitHub Actions vulnerabilities, enable CodeQL for GitHub Actions. For more information, see Configuring default setup for code scanning.

Opting out of built-in protections

If you have worked through the questions above and confirmed your workflow requires pull_request_target and uses it safely, you can opt out of the default event policy and actions/checkout protection.

Setting allow-unsafe-pr-checkout: true as an actions/checkout input allows checking out pull request head refs from forks. Only do this after confirming the checked-out code is never executed. The input is intentionally named to be easy to spot in code review and static analysis.

This protection only covers fork pull request refs. Checking out other untrusted code, such as an unrelated third-party repository, fetching code with git fetch or gh pr checkout, or running a downloaded artifact, is not covered by the actions/checkout checks.