Babfile Syntax

Define tasks in YAML format. Bab searches for Babfile, Babfile.yaml, or Babfile.yml.

Basic Task

tasks:
  build:
    desc: Build the application
    run:
      - cmd: go build -o app

Task Properties

desc - Description

Optional documentation for the task.

alias / aliases - Short Names

Define short aliases for tasks. Run bab t instead of bab test:

tasks:
  test:
    desc: Run tests
    alias: t
    run:
      - cmd: go test ./...

  build:
    desc: Build the application
    aliases: [b, bld]
    run:
      - cmd: go build -o app

Use alias for a single shortcut or aliases for multiple. Both can be combined:

tasks:
  deploy:
    alias: d
    aliases: [dep, ship]
    run:
      - cmd: ./deploy.sh

Aliases appear in bab --list and work with tab completion.

run - Commands

List of commands or task references to execute.

Shell Commands

Use the cmd key to run shell commands:

tasks:
  deploy:
    desc: Deploy application
    run:
      - cmd: npm test
      - cmd: npm run build
      - cmd: ./deploy.sh

Task References

Use the task key to run another task inline:

tasks:
  setup:
    run:
      - cmd: npm install

  build:
    run:
      - task: setup
      - cmd: npm run build

Task references support the same options as commands (silent, output, etc.).

deps - Dependencies

Tasks to run before this task.

tasks:
  setup:
    desc: Install dependencies
    run:
      - cmd: npm install

  build:
    desc: Build application
    deps: [setup]
    run:
      - cmd: npm run build

  deploy:
    desc: Deploy to production
    deps: [build, test]
    run:
      - cmd: ./deploy.sh

Namespaced Tasks

Use colon notation for task namespaces (flat structure, not nested YAML):

tasks:
  dev:start:
    desc: Start dev server
    run:
      - cmd: npm run dev

  dev:watch:
    desc: Watch files
    run:
      - cmd: npm run watch

  test:unit:
    desc: Unit tests
    run:
      - cmd: npm run test:unit

  test:e2e:
    desc: E2E tests
    run:
      - cmd: npm run test:e2e

Run with bab dev:start or bab test:unit.

Platform-Specific Commands

Run different commands based on the operating system using the platforms array:

tasks:
  deploy:
    desc: Deploy to production
    run:
      - cmd: ./scripts/deploy.sh
        platforms: [linux, darwin]
      - cmd: powershell scripts/deploy.ps1
        platforms: [windows]

Available platforms: linux, darwin, windows. Commands without a platforms array run on all platforms.

Conditional Execution

Skip tasks or run items based on conditions using the when key. Conditions support variable interpolation and comparison operators.

Task-Level Conditions

Skip an entire task if the condition is false:

tasks:
  deploy-prod:
    desc: Deploy to production
    when: ${{ environment }} == 'prod'
    run:
      - cmd: ./deploy.sh --env=prod

Run Item Conditions

Skip individual commands, task references, logs, or prompts:

tasks:
  build:
    run:
      - prompt: skip_tests
        type: confirm
        message: "Skip tests?"
        default: false
      - cmd: npm test
        when: ${{ skip_tests }} == 'false'
      - cmd: npm run build

Condition Syntax

Syntax	Description
${{ var }}	Truthy check - runs if variable is non-empty and not "false"
${{ var }} == 'value'	Equality - runs if variable equals value
${{ var }} != 'value'	Inequality - runs if variable does not equal value

Both single quotes ('value') and double quotes ("value") are supported.

Truthy Values

Values are evaluated as follows:

• Empty string ("") → falsy (skip)
• "false" (case-insensitive) → falsy (skip)
• Whitespace-only → falsy (skip)
• Undefined variables → falsy (skip)
• Any other value → truthy (run)

Using with Prompts

Conditions work with prompt results for dynamic workflows:

tasks:
  deploy:
    run:
      - prompt: confirm
        type: confirm
        message: "Deploy to production?"
        default: false
      - cmd: ./deploy.sh
        when: ${{ confirm }}
      - log: "Deployment skipped"
        when: ${{ confirm }} == 'false'

Complete Example

tasks:
  release:
    desc: Build and optionally deploy
    run:
      - prompt: environment
        type: select
        message: "Select environment:"
        options: [dev, staging, prod]
        default: dev

      - prompt: run_tests
        type: confirm
        message: "Run tests first?"
        default: true

      - cmd: npm test
        when: ${{ run_tests }}

      - cmd: npm run build

      - cmd: ./deploy.sh --env=dev
        when: ${{ environment }} == 'dev'

      - cmd: ./deploy.sh --env=staging
        when: ${{ environment }} == 'staging'

      - cmd: ./deploy.sh --env=prod
        when: ${{ environment }} == 'prod'

      - log: "Deployed to ${{ environment }}"

Silent Mode

The silent option suppresses command prompt display (e.g., $ echo hello). Useful for reducing noise when running many commands.

Global Silent

Apply to all tasks:

silent: true

tasks:
  build:
    run:
      - cmd: npm run build

Task Silent

Apply to a specific task:

tasks:
  install:
    silent: true
    run:
      - cmd: npm install

Command Silent

Apply to individual commands or task references:

tasks:
  build:
    run:
      - cmd: echo "Installing..."
        silent: true
      - cmd: npm install
      - task: test
        silent: true

Silent Precedence

Command-level overrides task-level, which overrides global. Default is false (show prompts).

Output Control

The output option controls whether stdout/stderr from commands is displayed. Different from silent which only affects the command prompt line.

Global Output

Apply to all tasks:

output: false

tasks:
  install:
    run:
      - cmd: npm install

Task Output

Apply to a specific task:

tasks:
  install:
    output: false
    run:
      - cmd: npm install

Command Output

Apply to individual commands or task references:

tasks:
  build:
    run:
      - cmd: npm install
        output: false
      - cmd: npm run build

Output Precedence

Command-level overrides task-level, which overrides global. Default is true (show output).

Working Directory

The dir option sets the working directory for command execution. Relative paths are resolved from the source Babfile's location.

Global Dir

Set the default working directory for all tasks:

dir: ./src

tasks:
  build:
    run:
      - cmd: npm run build  # runs in ./src

Task Dir

Override the global directory for a specific task:

dir: ./src

tasks:
  test:
    dir: ./tests
    run:
      - cmd: npm test  # runs in ./tests

Command Dir

Set the directory for a specific command:

tasks:
  deploy:
    run:
      - cmd: npm run build
        dir: ./frontend
      - cmd: go build
        dir: ./backend

Dir Precedence

Command-level overrides task-level, which overrides global. The default is the Babfile's directory.

dir: ./global

tasks:
  example:
    dir: ./task
    run:
      - cmd: pwd          # runs in ./task
      - cmd: pwd          # runs in ./cmd
        dir: ./cmd

Included Babfiles

Tasks from included Babfiles run in their source Babfile's directory by default. Relative paths in included tasks are resolved from the included Babfile's location, not the main Babfile.

# main/Babfile.yml
includes:
  api:
    babfile: ./api/Babfile.yml  # api tasks run in ./api by default

tasks:
  build:
    run:
      - cmd: pwd  # runs in ./main

Environment Variables

Define environment variables at three levels: global, task, or command. Variables cascade with lower levels overriding higher ones.

Global Environment

Set variables for all tasks at the root level:

env:
  NODE_ENV: production
  API_URL: https://api.example.com

tasks:
  build:
    run:
      - cmd: echo "Building for $NODE_ENV"

Task Environment

Set variables for a specific task:

tasks:
  dev:
    desc: Start development server
    env:
      PORT: "3000"
      DEBUG: "true"
    run:
      - cmd: npm run dev

Command Environment

Set variables for a specific command:

tasks:
  deploy:
    run:
      - cmd: ./deploy.sh
        env:
          DEPLOY_ENV: staging
      - cmd: ./notify.sh
        env:
          DEPLOY_ENV: production

Precedence

When the same variable is defined at multiple levels, command-level overrides task-level, which overrides global:

env:
  MODE: global

tasks:
  example:
    env:
      MODE: task
    run:
      - cmd: echo $MODE  # prints "command"
        env:
          MODE: command

Variables

Define reusable values with ${{ }} syntax. Variables are resolved by Bab before commands run.

Global Variables

vars:
  app_name: myapp
  version: "1.0.0"

tasks:
  build:
    run:
      - cmd: go build -o ${{ app_name }}

Task Variables

Override global variables within a task:

vars:
  mode: production

tasks:
  dev:
    vars:
      mode: development
    run:
      - cmd: echo "Running in ${{ mode }} mode"

Environment Access

Read OS environment variables with ${{ env.VAR }}:

vars:
  home: ${{ env.HOME }}
  target: ${{ env.GOOS }}

tasks:
  build:
    run:
      - cmd: echo "Building for ${{ target }}"

Variable References

Variables can reference other variables:

vars:
  base: /app
  build_dir: ${{ base }}/build
  output: ${{ build_dir }}/bin

tasks:
  build:
    run:
      - cmd: mkdir -p ${{ output }}

Export to Shell

Variables are not auto-exported. Use env: to pass to shell:

vars:
  app: myapp

tasks:
  run:
    env:
      APP_NAME: ${{ app }}
    run:
      - cmd: echo $APP_NAME

Escaping

Use $${{ to output literal ${{:

tasks:
  help:
    run:
      - cmd: echo "Use $${{ var }} syntax"

Interactive Prompts

Collect user input during task execution with the prompt run item. Prompt results are stored in variables for use in subsequent commands.

Prompt Types

Confirm (Yes/No)

tasks:
  deploy:
    run:
      - prompt: continue_deploy
        type: confirm
        message: "Deploy to production?"
        default: false
      - cmd: echo "Deploying..."

Result stored: "true" or "false"

Input (Free Text)

tasks:
  setup:
    run:
      - prompt: project_name
        type: input
        message: "Enter project name:"
        default: "my-app"
        placeholder: "name"
        validate: "^[a-z-]+$"  # optional regex
      - cmd: echo "Creating ${{ project_name }}"

Select (Single Choice)

tasks:
  configure:
    run:
      - prompt: environment
        type: select
        message: "Select environment:"
        options:
          - dev
          - staging
          - prod
        default: dev
      - cmd: echo "Selected: ${{ environment }}"

Multiselect (Multiple Choices)

tasks:
  install:
    run:
      - prompt: features
        type: multiselect
        message: "Select features to install:"
        options:
          - auth
          - api
          - ui
          - analytics
        defaults: [auth, api]
        min: 1
        max: 3
      - cmd: echo "Installing: ${{ features }}"

Result stored as comma-separated string: "auth,api,ui"

Password (Hidden Input)

tasks:
  login:
    run:
      - prompt: api_key
        type: password
        message: "Enter API key:"
        confirm: true  # require re-entry
      - cmd: ./auth.sh
        env:
          API_KEY: ${{ api_key }}

Number (Numeric Input)

tasks:
  scale:
    run:
      - prompt: replicas
        type: number
        message: "Number of replicas:"
        default: 3
        min: 1
        max: 10
      - cmd: kubectl scale --replicas=${{ replicas }}

Prompt Properties

Property	Type	Description
prompt	string	Variable name to store result (required)
type	string	Prompt type: confirm, input, select, multiselect, password, number (required)
message	string	Message to display (required)
default	string	Default value for confirm/input/select/number
defaults	array	Default selections for multiselect
options	array	Available choices for select/multiselect
placeholder	string	Placeholder text for input
validate	string	Regex pattern to validate input
min	integer	Minimum value (number) or selections (multiselect)
max	integer	Maximum value (number) or selections (multiselect)
confirm	boolean	Require password re-entry for confirmation
platforms	array	Run only on specified platforms
when	string	Condition to evaluate before running

Non-Interactive Mode (CI)

When running in non-interactive environments (no TTY), prompts use default values. If no default is set, the task fails with an error. Password prompts always fail in non-interactive mode.

Complete Example

vars:
  app_name: myapp

tasks:
  deploy:
    desc: Interactive deployment
    run:
      - prompt: confirm_deploy
        type: confirm
        message: "Deploy ${{ app_name }} to production?"
        default: false

      - prompt: environment
        type: select
        message: "Select target environment:"
        options: [staging, production]
        default: staging

      - prompt: replica_count
        type: number
        message: "Number of replicas:"
        default: 3
        min: 1
        max: 10

      - log: "Deploying to ${{ environment }} with ${{ replica_count }} replicas"
      - cmd: ./deploy.sh --env=${{ environment }} --replicas=${{ replica_count }}

Includes

Import tasks from other Babfiles with namespace prefixes:

includes:
  utils:
    babfile: ./tools/Babfile.yml

tasks:
  build:
    desc: Build everything
    deps: [utils:setup]
    run:
      - cmd: go build -o app

Tasks from the included file are prefixed with the namespace (e.g., utils:setup, utils:lint).

Complete Example

env:
  NODE_ENV: production

tasks:
  setup:
    desc: Install dependencies
    silent: true
    output: false
    run:
      - cmd: npm install

  lint:
    desc: Run linter
    deps: [setup]
    run:
      - cmd: npm run lint

  build:
    desc: Build application
    deps: [setup, lint]
    env:
      BUILD_MODE: release
    run:
      - cmd: npm run build

  test:
    desc: Run tests
    deps: [build]
    run:
      - cmd: npm test

  deploy:
    desc: Deploy to production
    deps: [build, test]
    run:
      - cmd: ./deploy.sh
        platforms: [linux, darwin]
        env:
          DEPLOY_ENV: production
      - cmd: powershell deploy.ps1
        platforms: [windows]
        env:
          DEPLOY_ENV: production