Laundry 👔

Air our your CloudFormation Templates.

CLI

laundry [template]

Commands:
  laundry lint [template]  lint a template

Options:
  --version  Show version number                                       [boolean]
  --help     Show help                                                 [boolean]

$ cat test.yaml
Resources: ''
$ laundry lint test.yaml
Root.Resources: must be an Object, got "" [RootValidator]

Library

import { lint } from 'laundry';

// type Error = {
//     path: Path;
//     message: string;
//     source?: string;
// };
// function lint(template: string, parameters?: object): Error[]

lint('{}');
// => [{ path: ['Root', 'Resources'], message: 'is required'}]

Better validation can be done if parameter values are provided:

lint('...', { ParamA: 1, ParamB: 'two'});

Ignoring Errors

Errors can be ignored via the LaundryIgnore Metadata property. The top-level Metadata property is specified as a map of paths within the CloudFormation template, written in dot notation. The * character can be used as a glob to match any value at that point in the path. For example, Resources.MyBucket.* would match all errors related to the MyBucket resource, and *.BucketName would ignore all errors related to properties called BucketName.

All LaundryIgnore properties in this example have the same effect:

Metadata:
  LaundryIgnore:
    # Match ResourcePropertyValidator errors for a specific property
    'Resources.A.Properties.Foo':
      - ResourcePropertyValidator
    # Match any ResourcePropertyValidator errors under "Resources.A"
    'Resources.A.*':
      - ResourcePropertyValidator
    # Match any ResourcePropertyValidator errors for properties named "Foo"
    '*.Foo':
      - ResourcePropertyValidator
    # Ignore all ResourcePropertyValidator errors
    '*':
      - ResourcePropertyValidator
Resources:
  A:
    Metadata:
      # Ignore ResourcePropertyValidator errors for this resource
      LaundryIgnore:
        - ResourcePropertyValidator
    Type: AWS::S3::Bucket
    Properties:
      Foo: bar

Development

Run tests using npm test or to do so in "watch" mode, use npm start.

Writing Additional Validators

Validation is done by "validators", which are visitors that are executed as Laundry walks the structure of the CloudFormation template. Validators are small in scope and typically assert a single property about a CloudFormation template.

Validators have the following structure and can implement any number of the following methods:

import { Validator } from './validate';

class SomethingValidator extends Validator {
  Root(path: Path, root: any): void { }
  AWSTemplateFormatVersion(path: Path, version: any): void { }
  Description(path: Path, description: any): void { }
  Metadata(path: Path, metadata: any): void { }
  Parameters(path: Path, parameters: any): void { }
  Mappings(path: Path, mappings: any): void { }
  Conditions(path: Path, conditions: any): void { }
  Transform(path: Path, transform: any): void { }
  Resources(path: Path, resources: any): void { }
  Outputs(path: Path, outputs: any): void { }
  Parameter(path: Path, parameter: any): void { }
  Mapping(path: Path, mapping: any): void { }
  Condition(path: Path, condition: any): void { }
  Resource(path: Path, resource: any): void { }
  CfnFn(path: Path, propertyName: string, value: CfnFn): void { }
  Output(path: Path, output: any): void { }
}

Existing validators can be seen in validators/.

Validation functions are available in the validate module for common validation needs. These functions are designed to be chained so that only one error message is created for each element. For example:

validate.required(value) && validate.string(value);

This will create only check if value is a string if it is considered to be present by the requried validation function. The functions return false when a violation is found. They can be used within a conditional to perform more complex validations, for example:

if (validate.required(value) && validate.string(value)) {
  if(_.includes([...], value)) ...
}

See validate for the full list of validation functions.

CloudFormation Resource Specification

The majority of validation rules come from the CloudFormation Resource Specification. This should be updated periodically using:

npm run update-resource-specification

Goals

The goal is to be as comprehensive as awslabs/cfn-python-lint but be available for use within Javascript-based projects or environments where Python is otherwise not available.