Restrictions when Dependabot triggers events
Dependabot is able to trigger GitHub Actions workflows on its pull requests and comments; however, certain events are treated differently.
For workflows initiated by Dependabot (github.actor == 'dependabot[bot]'
) using the pull_request
, pull_request_review
, pull_request_review_comment
, push
, create
, deployment
, and deployment_status
events, these restrictions apply:
GITHUB_TOKEN
has read-only permissions by default.- Secrets are populated from Dependabot secrets. GitHub Actions secrets are not available.
For workflows initiated by Dependabot (github.actor == 'dependabot[bot]'
) using the pull_request_target
event, if the base ref of the pull request was created by Dependabot (github.event.pull_request.user.login == 'dependabot[bot]'
), the GITHUB_TOKEN
will be read-only and secrets are not available.
These restrictions apply even if the workflow is re-run by a different actor.
For more information, see Keeping your GitHub Actions and workflows secure: Preventing pwn requests.
Troubleshooting failures when Dependabot triggers existing workflows
Après avoir configuré les mises à jour Dependabot pour GitHub.com, il se peut que constatiez des défaillances quand des workflows existants sont déclenchés par des événements Dependabot.
Par défaut, les exécutions de workflows GitHub Actions qui sont déclenchées par Dependabot à partir d’événements push
, pull_request
, pull_request_review
ou pull_request_review_comment
ont considérées comme ayant été ouvertes à partir d’une duplication (fork) de dépôt. Contrairement aux workflows déclenchés par d’autres acteurs, cela signifie qu’ils reçoivent un GITHUB_TOKEN
en lecture seule et qu’ils n’ont pas accès aux secrets qui sont normalement disponibles. Ainsi, les flux de travail qui tentent d’écrire dans le dépôt échouent quand ils sont déclenchés par Dependabot.
Il existe trois façons de résoudre ce problème :
- Vous pouvez mettre à jour vos workflows de sorte qu’ils ne soient plus déclenchés par Dependabot en utilisant une expression de ce type :
if: github.actor != 'dependabot[bot]'
. Pour plus d’informations, consultez « Évaluer les expressions dans les workflows et les actions ». - Vous pouvez modifier vos workflows de sorte qu’ils utilisent un processus en deux étapes qui comprend
pull_request_target
qui ne présente pas ces limites. Pour plus d’informations, consultez « Automatisation de Dependabot avec GitHub Actions ». - Vous pouvez permettre aux workflows déclenchés par Dependabot d’accéder aux secrets et autoriser le terme
permissions
à augmenter l’étendue par défaut deGITHUB_TOKEN
.
Some troubleshooting advice is provided in this article. You can also see Workflow syntax for GitHub Actions.
Accessing secrets
When a Dependabot event triggers a workflow, the only secrets available to the workflow are Dependabot secrets. GitHub Actions secrets are not available. You must therefore store any secrets that are used by a workflow triggered by Dependabot events as Dependabot secrets. For more information, see Configuration de l’accès aux registres privés pour Dependabot.
Dependabot secrets are added to the secrets
context and referenced using exactly the same syntax as secrets for GitHub Actions. For more information, see Utilisation de secrets dans GitHub Actions.
If you have a workflow that will be triggered by Dependabot and also by other actors, the simplest solution is to store the token with the permissions required in an action and in a Dependabot secret with identical names. Then the workflow can include a single call to these secrets. If the secret for Dependabot has a different name, use conditions to specify the correct secrets for different actors to use.
For examples that use conditions, see Automatisation de Dependabot avec GitHub Actions.
To access a private container registry on AWS with a user name and password, a workflow must include a secret for username
and password
.
In this example, when Dependabot triggers the workflow, the Dependabot secrets with the names READONLY_AWS_ACCESS_KEY_ID
and READONLY_AWS_ACCESS_KEY
are used. If another actor triggers the workflow, the actions secrets with those names are used.
name: CI on: pull_request: branches: [ main ] jobs: build: runs-on: ubuntu-latest steps: - name: Checkout repository uses: actions/checkout@v4 - name: Login to private container registry for dependencies uses: docker/login-action@3b4c5d6e7f8a9b0c1d2e3f4a5b6c7d8e9f0a1b2c with: registry: https://1234567890.dkr.ecr.us-east-1.amazonaws.com username: ${{ secrets.READONLY_AWS_ACCESS_KEY_ID }} password: ${{ secrets.READONLY_AWS_ACCESS_KEY }} - name: Build the Docker image run: docker build . --file Dockerfile --tag my-image-name:$(date +%s)
name: CI
on:
pull_request:
branches: [ main ]
jobs:
build:
runs-on: ubuntu-latest
steps:
- name: Checkout repository
uses: actions/checkout@v4
- name: Login to private container registry for dependencies
uses: docker/login-action@3b4c5d6e7f8a9b0c1d2e3f4a5b6c7d8e9f0a1b2c
with:
registry: https://1234567890.dkr.ecr.us-east-1.amazonaws.com
username: ${{ secrets.READONLY_AWS_ACCESS_KEY_ID }}
password: ${{ secrets.READONLY_AWS_ACCESS_KEY }}
- name: Build the Docker image
run: docker build . --file Dockerfile --tag my-image-name:$(date +%s)
Changing GITHUB_TOKEN
permissions
By default, GitHub Actions workflows triggered by Dependabot get a GITHUB_TOKEN
with read-only permissions. You can use the permissions
key in your workflow to increase the access for the token:
name: CI on: pull_request # Set the access for individual scopes, or use permissions: write-all permissions: pull-requests: write issues: write repository-projects: write ... jobs: ...
name: CI
on: pull_request
# Set the access for individual scopes, or use permissions: write-all
permissions:
pull-requests: write
issues: write
repository-projects: write
...
jobs:
...
For more information, see Authentification par jeton automatique.
Manually re-running a workflow
When you manually re-run a Dependabot workflow, it will run with the same privileges as before even if the user who initiated the rerun has different privileges. For more information, see Ré-exécution de workflows et de travaux.