Pub autobot pub score pipeline/ci release status pipeline/ci delevop status dart supported platforms dart sdk version MIT-license

Autobot is a command-line templating engine for automating standardised repetitive operations, such as creating a new project with a specific folder structure or creating files and classes according to a predefined scheme to ensure compliance with architectural specifications. Autobot can also be used for configuring projects by creating build configurations both locally and in the CI.


  1. Install Dart following these instructions.

  2. Activate autobot running the following command:

$ dart pub global activate autobot

Check for the autobot version to see if autobot was successfully installed:

$ autobot version


Autobot needs a .autobot_config.yaml file. Create a local or global one via the following commands:

# creates local config file in current working directory
$ autobot init
# creates local config file in given path (-p)
$ autobot init -p subfolder/subsubfolder/
# creates global config file in home directory
$ autobot init -g

This is a .autobot_config.yaml with all available configuration.

  # tell autobot where is should search for tasks
  taskDir: some/relative_path/to_tasks_directory/


Run an autobot task via the following command:

# assuming there is a my_task.yaml inside taskDir
$ autobot run --task my_task
# use -t instead of --task if you like short cuts :)

You can also omit run --task if you want:

$ autobot my_task

Autobot tasks are YAML files witch describe what needs to be done.

Pass flags to a task using : as prefix like so:

$ autobot my_task:first_flag:second_flag

Autobot assigns the flags to keys with the following schema flag<index>: This means that the first flag will be assigned to flag1, the second will be assigned to flag2 and so on.

You can add some meta data like name and description to a task like this:

  name: My task
  description: This is my autobot task

Use steps to define the actuall work like this:

  name: My task
  description: This is my autobot task

  - read:
    file: some/path/to/file.yaml

  - ask:
    key: name
    prompt: Whats your name?

  - ask:
    key: age
    prompt: How old are you?

  - write:
    file: some/path/output_file.txt
    content: My name is {{name}} and I'm {{age}} years old.

Available steps

Check out all available steps. Keep in mind that all fields of all steps (except vars) can be mustache templates for dynamic customization. Check out the Examples for more details.


Asks the user for input and assigns that input to the key.

- ask:
  key: # String -> the key the value should be asssigned to
  prompt: # String -> the prompt to print


Runs a shell command.

- command:
  run: # String -> shell command


Runs a javascript. NOTE: Requires that node is installen in your commond line tool.

- javascript:
  run: # String -> javascript


Reads a yaml file. The data can then be used in other spets like write or javascript.

- read:
  file: # String -> path of yaml file


Runs a sub task.

- runTask:
  file: # String -> name of that which is loacted in taskDir of .autobot_config.yaml


Define some data that can be used in other steps like write or javascript.

- vars:
  key: value # some key value pair
  dataArray: [a, b c] # some array
    key: value
    key2: value2
    - a
    - b
    - c
  someText: |
    Lorem ipsum...


Writes a file.

- write:

  # The path of the file which should be written

  # Whether writting is enabled or not.
  enabled: # n, no, false or y, yes, true

  # How the file should be written
  writeMethod: # keepExistingFile (default), replaceExistingFile, extendFile

  # Works only if writeMethod is extendFile
  extendAt: # Values: top, bottom, custom regex.

  # The content of the file


Autobot automatically reads all variables from the environment and provides them to all steps. This is very helpful if you want autobot to use variables from your CI or if you want integrate autobot into so existing processes or cli commands.

Autobot and CI

Since a CI doesn't support interactive scripts, autobot offers the possibility to set input values via arguments using -i or --input:

$ autobot run --task some_autobot_task --input userName=somename,userAge=23

This will skip the interactive command line prompt for all set values.


Javascript example

  - aks:
    key: userName
    prompt: What is your name?

  - ask:
    key: userAge
    prompt: How old are you?

  - javascript:
    run: |
      // Access the value of userName
      var userName = autobot.variables.userName
      // Check whether userName is blank. If yes, set a fallback value.
      if (!userName || userName.length === 0) {
        autobot.variables.userName = 'Some fallback name'
      // Define a new variable
      autobot.variables.varFormJs = 'Hello!'

  - write:
    file: some/relative/path.txt
    content: Hi {{userName}}

  - write:
    file: /some/non_relative/path.txt
    content: |
      Hi {{userName}},
      is it true that you are {{userAge}} years old?
      This is a value from javascript: {{varFormJs}}

This example task ask the user for his name and age and will create two files. The first file will be $pwd/some/relative/path.txt with the following content:

Hi Peter

The second file will be /some/non_relative/path.txt with the following content:

Hi Peter,
is it true that you are 30 years old?
This is a value from javascript: Hello!

Read example

Assuming there is a yaml file like this on your machine called task_data.yaml:

key1: value1
key2: value2

  mapKey1: mapValue1
  mapKey2: mapValue2

  - key: a
  - key: b
  - key: c

The task:

  - read:
    file: task_data.yaml

  - write:
    file: output.txt
    content: |
      Pairs: {{key1}}, {{key2}}
      Map pairs: {{varMap.mapKey1}} {{varMap.mapKey2}}
        -> {{key}} <-

Tis task will read the task_data.yaml and create the following file:

Pairs: value1, value2
Map pairs: mapValue1 mapValue2
  -> a <-
  -> b <-
  -> c <-