Skip to content

Manage your Cloudflare credentials, securely

License

Notifications You must be signed in to change notification settings

jacobbednarz/cf-vault

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

cf-vault

Manage your Cloudflare credentials, securely.

The goal of this project is to ensure that when you need to interact with Cloudflare:

  • You're not storing credentials unencrypted; and
  • You're not exposing your credentials to the entire environment or to all processes; and
  • You're not using long lived credentials

To achieve this, cf-vault uses the concept of profiles with associated scopes to either generate short lived API tokens or retrieve the API key from secure storage (such as Mac OS keychain).

Demo

demo

Install

$ brew tap jacobbednarz/tap
$ brew install jacobbednarz/tap/cf-vault

Getting started

  1. First step is to generate a new API key or API token. Either will work however there are some subtle differences to take into consideration before choosing your path.

    • API tokens are not supported by all services yet. Regardless of whether you are using the short lived credentials or long lived token, it may not work for all services and you may need to have a backup profile defined using an API key to cover all scenarios .
    • API keys are supported everywhere however they cannot be scoped. API keys have the permission and scopes that your user account has. This can be dangerous so be sure to tread carefully as it may have unintended consequences.

    While it is possible (and better practice of principle of least privilege), to use an API token with only permissions to create a new API token, this isn't really viable for all use cases yet. The recommended approach is to use the API key for the profile and rely on a custom policy to scope the short lived credential. This allows the best of both worlds where if you need to use a service that doesn't support API tokens, you don't need to create a new profile.

    To create a new API token:

    > https://dash.cloudflare.com/
      > My Profile
        > API Tokens
          > Create API token
    

    To retrieve your API key:

    > https://dash.cloudflare.com/
      > My Profile
        > API Tokens
          > Global API Key
    
  2. If you're using an API key, you can skip to the next step. Otherwise, navigate through the UI and configure what permissions and resources you'd like to assign to the token. If you're looking to use an API token to generate short lived API tokens, you should only need the single predefined "Create API tokens" permission. See the section below on generating the desired TOML output for instructions on how to do automatically convert policies from API responses.

    Note: Be sure to note down the API token before closing/navigating away from the UI as you won't be able to retrieve it again.

  3. Once you have your API key or API token value, you can start using cf-vault by creating a profile. A profile is the collection of configuration that tells cf-vault how you intend to interact with the Cloudflare credentials. You need to start by calling cf-vault add [your-profile-name] where [your-profile-name] is a label for what the credential/use of the profile is. Some examples:

    • cf-vault add write-everything
    • cf-vault add read-only
    • cf-vault add super-scary-access-everything
    • cf-vault add api-token-to-create-other-tokens

    There is no limit on how many profiles you have if you prefer to have specific profiles for your use cases.

  4. Now that you have created a profile, you can use it with cf-vault exec [your-profile-name].

If you do not wish to use the short lived credentials functionality, that's totally fine and you can do so by omitting the session_duration value and instead the long lived credentials you've setup will be used.

Usage

cf-vault allows you to manage your Cloudflare credentials in a safe place and only expose the credentials to the processes that require them and only for a limited timespan.

$ env | grep -i cloudflare
# => no results

$ cf-vault exec work -- env | grep -i cloudflare
CLOUDFLARE_VAULT_SESSION=work
[email protected]
CLOUDFLARE_API_KEY=s3cr3t
[email protected]
CF_API_KEY=s3cr3t

If you don't provide a command, you will be dropped into a new shell with the credentials populated.

$ cf-vault exec work
$ env | grep -i cloudflare
CLOUDFLARE_VAULT_SESSION=work
[email protected]
CLOUDFLARE_API_KEY=s3cr3t
[email protected]
CF_API_KEY=s3cr3t

$ exit
$ env | grep -i cloudflare
# => no results

Predefined short lived token policies

If you don't need to generate a custom token policy, you can instead use one of the predefined templates which takes care of the heavy lifting for you. You can use read-only (read all resources) or write-everything (write all resources) as the --profile-template flag and it will generate everything needed behind the scenes on your behalf. Note: You still need to provide --session-duration as well otherwise the short lived tokens will not be generated.

Examples:

  • cf-vault add my-read-profile-name --profile-template "read-only" --session-duration "15m"
  • cf-vault add my-write-profile-name --profile-template "write-everything" --session-duration "15m"

Generating token policies

While TOML is more readable, its not always straight forward to generate the desired output. Instead, you can use the Cloudflare dashboard to build the policy you'd like and then covert that to TOML using some tooling to avoid manually building your policy (though you can if you understand the syntax!).

  1. Using cf-vault add create your profile following the prompts.

  2. Create the token you'd like to use on the command line using the Cloudflare dashboard.

  3. Make the API call to fetch the token you've just created. See https://developers.cloudflare.com/api/operations/user-api-tokens-token-details or https://developers.cloudflare.com/api/operations/user-api-tokens-list-tokens to fetch all tokens.

  4. Write the contents of the single result JSON payload to a local file. For the example, I'll use example_token.json for the documentation.

  5. Run the following command using docker which will pull the go-toml tool for coverting JSON -> TOML. Remember to replace example_token.json with your filename.

    docker run -v $PWD:/workdir pelletier/go-toml jsontoml /workdir/example_token.json
    
  6. Paste the generated policy into your configuration file. You will need to adjust the structure slightly to match the hierarchy. For instance, if I have the following profile:

[profiles]

[profiles.doco-example]
  auth_type = "api_token"
  email = "[email protected]"
  session_duration = "15m"

and my policy output:

[[policies]]
effect = "allow"

[[policies.permission_groups]]
  id = "eb258a38ea634c86a0c89da6b27cb6b6"
  name = "Access: Apps and Policies Read"

[[policies.permission_groups]]
  id = "517b21aee92c4d89936c976ba6e4be55"
  name = "Zone Settings Read"

[[policies.permission_groups]]
  id = "c8fed203ed3043cba015a93ad1616f1f"
  name = "Zone Read"

# .. snip

The policy needs to be updated to prepend profiles.doco-example to the section keys.

[[profiles.doco-example.policies]]
effect = "allow"

[[profiles.doco-example.policies.permission_groups]]
  id = "eb258a38ea634c86a0c89da6b27cb6b6"
  name = "Access: Apps and Policies Read"

[[profiles.doco-example.policies.permission_groups]]
  id = "517b21aee92c4d89936c976ba6e4be55"
  name = "Zone Settings Read"

[[profiles.doco-example.policies.permission_groups]]
  id = "c8fed203ed3043cba015a93ad1616f1f"
  name = "Zone Read"

# .. snip

Making the complete configuration look like:

  [profiles]

  [profiles.doco-example]
    auth_type = "api_token"
    email = "[email protected]"
    session_duration = "15m"

    [[policies]]
    effect = "allow"

    [[policies.permission_groups]]
      id = "eb258a38ea634c86a0c89da6b27cb6b6"
      name = "Access: Apps and Policies Read"

    [[policies.permission_groups]]
      id = "517b21aee92c4d89936c976ba6e4be55"
      name = "Zone Settings Read"

    [[policies.permission_groups]]
      id = "c8fed203ed3043cba015a93ad1616f1f"
      name = "Zone Read"

    # .. snip