This action will check that all checks have passed on a given pull request.
- Upgrade to Node24 runtime
- On GitHub.com, the name of the check run created by the workflow itself is extracted from the available
job.check_run_idcontext, using this action as a reusable is now supported. This functionality is not yet available on GitHub Enterprise Server (GHES), will likely be available on the next release version - 3.18. On GitHub Enterprise Server, the action will attempt to extract its check name from the workflow file. - Support for commit statuses is now implemented on V2. To prevent a breaking change, this is disabled by default. You can enable it by setting the
include_commit_statusesinput totrue.
The action works well with the pull_request or pull_request_target events. You can use it to confirm that all checks (and optionally commit statuses) that run on a pull request passed.
name: All checks pass
on:
pull_request:
types: [opened, synchronize, reopened, ready_for_review]
jobs:
allchecks:
runs-on: ubuntu-latest
permissions:
checks: read
contents: read
steps:
- uses: wechuli/allcheckspassed@v2With this default configuration, the action will fail if any of the checks on the pull request have failed or if any of the checks are still in progress, pending or queued when the workflow is complete.
The action also created a checks summary with details of each check that was evaluated and their status:
When the workflow is triggered, the action will poll the GitHub API every 1 minute (default) for 10 tries (default) - you can change these to your liking. If all the checks are successful, the action will succeed. If any of the checks are still in progress, pending or queued, the action fails
It will create a job summary of each check along with the details.
The workflow job must be granted read access to checks and contents for it to work:
jobs:
allchecks:
runs-on: ubuntu-latest
permissions:
checks: read
contents: readIf you additionally want the action to check for commit statues on top of checks, you'll need to add the statuses permission:
jobs:
allchecks:
runs-on: ubuntu-latest
permissions:
checks: read
contents: read
statuses: readIf you enable the ignore_superseded_runs option (see Ignore superseded workflow runs below), the action additionally calls the GitHub Actions Runs API, which requires the actions: read permission:
jobs:
allchecks:
runs-on: ubuntu-latest
permissions:
checks: read
contents: read
actions: readYou can exclude certain checks from causing a failure by using the checks_exclude input. The input accepts a comma
separated string of values:
steps:
- uses: wechuli/allcheckspassed@v2
with:
checks_exclude: "CodeQL,lint,cosmetic"You might want to do this to allow certain checks to fail, such as a code quality check, but still require that all other checks pass.
The checks_exclude and checks_include inputs support Regular Expressions. For example, if you want to exclude all
checks have the word lint somewhere in the string, you don't have to list them all out, you can use the following:
steps:
- uses: wechuli/allcheckspassed@v2
with:
checks_exclude: ".*lint.*"You can choose to include only specific checks for evaluation and ignore others. This is not the primary use case
for this action, but it is possible. If you want the check to always be included, you might consider using the native
repository rulesets or branch protection rules. You can't provide both the checks_include and checks_excludeinputs.
steps:
- uses: wechuli/allcheckspassed@v2
with:
checks_include: "CodeQL"At the moment, checks with success, neutral and skipped are considered passing by GitHub. This action will
default to this behavior as well but you can change this if you want to.
steps:
- uses: wechuli/allcheckspassed@v2
with:
treat_skipped_as_passed: false
treat_neutral_as_passed: falseIn the above configuration, the action will fail if any of the checks are skipped or neutral.
If you define one or more checks in the checks_include input, the action will output a warning if any of the checks
are missing. You can choose instead to fail the action if any of the checks are missing:
steps:
- uses: wechuli/allcheckspassed@v2
with:
fail_on_missing_checks: trueThe default behavior of this action is that is delays its execution for 1 minute (default) and polls the API 10 times ( retries) with a delay between each poll of 1 minute (default). You can change these values to your liking:
steps:
- uses: wechuli/allcheckspassed@v2
with:
delay: "5"
retries: "3"
polling_interval: "3"When the step with the action runs, it will wait for 5 minutes before polling the API for the first time. It will then poll the API 3 times with a delay of 3 minutes between each poll.
You also don't have to poll, you can just run the action once and get the result.
steps:
- uses: wechuli/allcheckspassed@v2
with:
poll: falseBy default, the action will fail the step as soon as a check meets the condition for failure. This default behavior can save some execution minutes since the action will not wait for all checks to complete before it reports a failure.The fail_fast flag can be used to cause this action to report a failure status as soon as one other check has failed. If you instead wish to wait for all checks to complete before reporting a failure, you can set the fail_fast flag to false.
steps:
- uses: wechuli/allcheckspassed@v2
with:
fail_fast: falseTo enable additional logging when the action runs, you may enable the verbose mode (defaults to false). The additional
logs will indicate which specific checks are being waited on in each polling iteration. This may be helpful in debugging
what checks are matched by checks_include or checks_exclude.
steps:
- uses: wechuli/allcheckspassed@v2
with:
verbose: trueBy default, the action will create a job summary with the details of each check that was evaluated and their status. You can disable the job summary by setting the show_job_summary input to false. This will prevent the action from creating a job summary at the end of the workflow run.
steps:
- uses: wechuli/allcheckspassed@v2
with:
show_job_summary: falseYou can choose to include commit statuses in addition to checks for evaluation. By default, this is disabled. If your CI/CD is only ever running on GitHub Actions and you are not explicitly calling the commit status API, then that option can be left to default to false as is. If you have an external integration (such as Jenkins) that is reporting commit statuses instead of checks, you can set the include_commit_statuses option to true so that those are evaluated as well.
steps:
- uses: wechuli/allcheckspassed@v2
with:
include_commit_statuses: trueChecks and status commits use a different API and there are some nuances to be aware of. More on this is documented at ./docs/adrs/commit_status.md
By default, the action evaluates every check run that the Checks API returns for the commit and deduplicates by (check name, app id). This means that if a GitHub Actions workflow is re-run on the same commit, stale check runs from the older run may still be evaluated when their names do not exactly match the new run (for example, matrix jobs in a cancelled run whose templates were never expanded). See #104 for a real-world example.
You can opt in to filtering check runs down to the latest GitHub Actions run per workflow on the commit by setting ignore_superseded_runs to true:
steps:
- uses: wechuli/allcheckspassed@v2
with:
ignore_superseded_runs: trueWhen this is enabled, the action calls GET /repos/:owner/:repo/actions/runs?head_sha=<sha> in addition to the Checks API, so the workflow must grant the actions: read permission:
jobs:
allchecks:
runs-on: ubuntu-latest
permissions:
checks: read
contents: read
actions: read
steps:
- uses: wechuli/allcheckspassed@v2
with:
ignore_superseded_runs: trueThis mode and the default raw-checks mode behave differently in some edge cases (notably around manually-created check runs that live in cancelled check suites). The two modes, what each one does, and guidance on when to use which are documented in ./docs/adrs/evaluation_modes.md.
This action is essentially a workflow run that will consume your GitHub Actions minutes. You may want to delay the execution of the action to give enough time for your checks to run. GitHub provides the environments feature that has a protection rule to allow you to delay the execution of job for a specified amount of time after the job is triggered:
This will save you some minutes if have checks that take a long time to complete and you don't want to poll the API for a long time.
jobs:
allchecks:
runs-on: ubuntu-latest
environment:
name: delayenv
deployment: false
steps:
- uses: wechuli/allcheckspassed@v2
with:
poll: falseUnfortunately, this feature is only available on private/internal repositories for Enterprise plans. All public repositories have access to this feature.
You want to require the check that is created to always pass in your repository rulesets or branch protection rules. Where possible prefer to configure repository rulesets to branch protection rules as they are more flexible.
Ultimately this is a temporary workaround for a missing feature, ensure all checks that run pass. GitHub may implement this as some point, at which point you will not need the action anymore.
- Unfortunately, you'll need to poll the API to get the state of the checks. The action itself is consuming GitHub Actions minutes doing this polling

