The sensitive connection details, such as the access token used by JFrog CLI on the JFrog platform, can be automatically generated by the action instead of storing it as a secret in GitHub. This is made possible by leveraging the OpenID-Connect (OIDC) protocol. This protocol can authenticate the workflow issuer and supply a valid access token, requiring only the JF_URL environment variable. Learn more about this integration in this blog post. To utilize the OIDC protocol, follow these steps:

1. Configure an OIDC Integration: This phase sets an integration between GitHub Actions to the JFrog platform.
   1. Navigate to the Administration tab In the JFrog Platform UI
   2. Click General | Manage Integrations
   3. Click New Integration | OpenID Connect:
   4. Configure the OIDC integration:

2. Configure an identity mapping: This phase sets an integration between a particular GitHub repository to the JFrog platform.

   An identity mapping is a configuration object utilized by the JFrog Platform to associate incoming OIDC claims with particular selected fields. These fields might include repository, actor, workflow, and others. To configure the identity mapping, click on the identity mapping created in section 1 and then click on Add Identity Mapping. Fill in priority 1 and fill out all required fields:

   You have the flexibility to define any valid list of claims required for request authentication. You can check a list of the possible claims here. Example Claims JSON:

   {
       "repository": "repository-owner/repository-name"
   }

1. Set required permissions: In the course of the protocol's execution, it's imperative to acquire a JSON Web Token (JWT) from GitHub's OIDC provider. To request this token, it's essential to configure the specified permission in the workflow file:

   permissions:
       id-token: write

2. Pass the 'oidc-provider-name' input to the Action (Required): The 'oidc-provider-name' parameter designates the OIDC configuration whose one of its identity mapping should align with the generated JWT claims. This input needs to align with the 'Provider Name' value established within the OIDC configuration in the JFrog Platform.

3. Pass the 'oidc-audience' input to the Action (Optional): The 'oidc-audience' input defines the intended recipients of an ID token (JWT), ensuring access is restricted to authorized recipients for the JFrog Platform. By default, it contains the URL of the GitHub repository owner. It enforces a condition, allowing only workflows within the designated repository/organization to request an access token. Read more about it here.

Example step utilizing OpenID Connect:

- uses: jfrog/setup-jfrog-cli@v4
  env:
      JF_URL: ${{ vars.JF_URL }}
  with:
      oidc-provider-name: setup-jfrog-cli

Notice: When using OIDC authentication, this action outputs both the OIDC token and the OIDC token username. These can be utilized within the current workflow to log into the JFrog platform through other actions or clients (e.g., for use with docker login). The added outputs are oidc-token and oidc-user, respectively.

The connection details of the JFrog platform used by this action can be stored as GitHub secrets (or GitHub Variables for non-secret values)

You can set the connection details to your JFrog Platform by using one of the following combinations:

1. JF_URL (no authentication)
2. JF_URL + JF_USER + JF_PASSWORD (basic authentication)
3. JF_URL + JF_ACCESS_TOKEN (authentication using a JFrog Access Token)

You can use these environment variables in your workflow as follows:

- uses: jfrog/setup-jfrog-cli@v4
  env:
    # JFrog Platform url
    JF_URL: ${{ vars.JF_URL }} # or 'https://acme.jfrog.io'

    # Basic authentication credentials
    JF_USER: ${{ secrets.JF_USER }}
    JF_PASSWORD: ${{ secrets.JF_PASSWORD }}
    # or
    # JFrog Platform access token
    JF_ACCESS_TOKEN: ${{ secrets.JF_ACCESS_TOKEN }}
- run: |
    jf rt ping

1. Make sure JFrog CLI is installed on your local machine by running jf -v.
2. Configure the details of the JFrog platform by running jf c add.
3. Export the details of the JFrog platform you configured, using the server ID you chose. Do this by running jf c export <SERVER ID>.
4. Copy the generated Config Token to the clipboard and save it as a secret on GitHub.

To use the saved JFrog platform configuration in the workflow, all you need to do it to expose the secret to the workflow. The secret should be exposed as an environment variable with the JFENV prefix. Here's how you do this:

- uses: jfrog/setup-jfrog-cli@v4
  env:
      JF_ENV_1: ${{ secrets.JF_SECRET_ENV_1 }}
- run: |
      # Ping the server
      jf rt ping

As you can see in the example above, we created a secret named JF_SECRET_ENV_1 and exposed it to the workflow as the JF_ENV_1 environment variable. That's it - the ping command will now ping the configured Artifactory server.

If you have multiple Config Tokens as secrets, you can use all of them in the workflow as follows:

- uses: jfrog/setup-jfrog-cli@v4
  env:
      JF_ENV_1: ${{ secrets.JF_SECRET_ENV_1 }}
      JF_ENV_2: ${{ secrets.JF_SECRET_ENV_2 }}
- run: |
      # Set the utilized JFrog configuration by providing the server ID (configured by the 'jf c add' command).
      jf c use local-1
      # Ping local-1 Artifactory server
      jf rt ping
      # Now use the second sever configuration exposed to the Action.
      jf c use local-2
      # Ping local-2 Artifactory server
      jf rt ping

Build info collection and publication to Artifactory happens seamlessly when using the action in your workflow.
The Action automatically sets the following environment variables: JFROG_CLI_BUILD_NAME and JFROG_CLI_BUILD_NUMBER with the workflow name and run number respectively. You therefore don't need to specify the build name and build number on any of the build related JFrog CLI commands.

In the following example, all downloaded files are registered as dependencies of the build and all uploaded files are registered as the build artifacts.

- run: |
      jf rt download "my-repo/artifacts/*" local-dir/
      jf rt upload "local-dir/*" new-repo/my-artifacts/
      jf rt build-publish

You may override the default build name and number by setting the following environment variables in your workflow:

env:
  JFROG_CLI_BUILD_NAME: "Custom build name"
  JFROG_CLI_BUILD_NUMBER: "123"

By default, the JFrog CLI version set in action.yml is used. To set a specific version, add the version input as follows:

- uses: jfrog/setup-jfrog-cli@v4
  with:
      version: X.Y.Z

It is also possible to set the latest JFrog CLI version by adding the version input as follows:

- uses: jfrog/setup-jfrog-cli@v4
  with:
      version: latest

JFrog Projects are a management entity that hosts all your resources related to a specific project, such as repositories, builds and Release Bundles.

By default, the JFrog CLI accepts a project flag in some of its commands. You can set the project key in the environment variable JF_PROJECT to avoid passing it in each command.

- uses: jfrog/setup-jfrog-cli@v4
  env:
      JF_PROJECT: "project-key"

If your agent has no Internet access, you can configure the workflow to download JFrog CLI from a remote repository in your JFrog Artifactory, which is configured to proxy the official download URL.

Here's how you do this:

1. Create a remote repository in Artifactory. Name the repository jfrog-cli-remote and set its URL to https://releases.jfrog.io/artifactory/jfrog-cli/

2. Set download-repository input to jfrog-cli-remote:

   - uses: jfrog/setup-jfrog-cli@v4
     env:
         # JFrog platform url (for example: https://acme.jfrog.io)
         JF_URL: ${{ vars.JF_URL }}
         JF_ACCESS_TOKEN: ${{ secrets.JF_ACCESS_TOKEN }}
   
     with:
         download-repository: jfrog-cli-remote