Skip to main content



The default Goodcheck configuration file is named goodcheck.yml. An example of the configuration is like the following:

rules:  - id: com.example.github    pattern: Github    severity: warning    message: |      GitHub is GitHub, not Github
      You may misspelling the name of the service!    justification:      - When you mean a service different from GitHub      - When GitHub is renamed    glob:      - app/views/**/*.html.slim      - config/locales/**/*.yml    pass:      - <a>Signup via GitHub</a>    fail:      - <a>Signup via Github</a>
import:  - goodcheck/*.yml
exclude:  - vendor

A rule hash under the rules list contains the following attributes:

idA string to identify a ruleyes
messageA message to tell writers why the code piece should be revisedyes
patternA pattern or patterns of text to be scannedno
justificationMessages to tell writers when an exception can be allowedno
globA glob or globs of files to be scannedno
severityA severity of a ruleno
passA pattern or patterns that do not match this ruleno
failA pattern or patterns that match this ruleno


The pattern can be one of either:

string literal#

A string literal represents a literal pattern or regexp pattern.

pattern:  - This is a literal pattern  - /This is a regexp pattern/  - /This is a regexp pattern with the case-insensitive option/i  - /This is a regexp pattern with the multiline option/m

If the string value begins with / and ends with /, it is a regexp pattern. You can optionally specify regexp options like /case-insensitive/i or /multiline/m.

literal pattern#

A literal pattern allows you to construct a regexp which matches exactly to the literal string.

id: com.sample.GitHubpattern:  literal: Github  case_sensitive: truemessage: Write GitHub, not Github

All regexp meta characters included in the literal value will be escaped. case_sensitive is an optional attribute and the default is true.

regexp pattern#

A regexp pattern allows you to write a regexp with meta characters.

id: com.sample.digitspattern:  regexp: \d{4,}  case_sensitive: false  multiline: falsemessage: Insert delimiters when writing large numbersjustification:  - When you are not writing numbers, including phone numbers, zip code, ...

It accepts two optional attributes, case_sensitive and multiline. The default values of case_sensitive and multiline are true and false respectively.

The regexp will be passed to Regexp.compile of Ruby. The precise definition of regular expressions can be found in the documentation for Ruby.

token pattern#

A token pattern compiles to a tokenized regexp.

id:  token: "<blink"  case_sensitive: falsemessage: Stop using <blink> tagglob: "**/*.html"justification:  - If Lynx is the major target of the web site

It tries to tokenize the input and generates a regexp which matches a sequence of tokens. The tokenization is heuristic and may not work well for your programming language. In that case, try using a regexp pattern.

The generated regexp of <blink is <\s*blink\b/m. It matches with <blink /> and < BLINK>, but does not match with

It accepts one optional attribute case_sensitive. The default value of case_sensitive is true. Note that the generated regexp is in multiline mode.

A token pattern can have an optional where attribute and variable bindings.

pattern:  - token: bgcolor=${color:string}    where:      color: true

The variable binding consists of a variable name and a variable type, where color and string in the example above respectively. You have to add a key of the variable name in where attribute.

Goodcheck has 8 built-in patterns:

  • string
  • int
  • float
  • number
  • url
  • email
  • word
  • identifier

You can find the exact definitions of the types in the definition of Goodcheck::Pattern::Token (@@TYPES).

You can omit the type of variable binding.

pattern:  - token: "margin-left: ${size}px;"    where:      size: true  - token: "backgroundColor={${color}}"    where:      color: true

In this case, the following character will be used to detect the range of binding. In the first example above, the px will be used as the marker for the end of size binding.

If parens or brackets are surrounding the variable, Goodcheck tries to match with nested ones in the variable. It expands five levels of nesting. See the example of matches with the second backgroundColor pattern:

  • backgroundColor={color} matches (color=="color")
  • backgroundColor={{ red: red(), green: green(), blue: green()-1 }} matches (color=="{ red: red(), green: green(), blue: green()-1 }")
  • backgroundColor={ {{{{{{}}}}}} } matches (color==" {{{{{{}}}}}")


The glob can be a string or a hash.

glob: "**/test/**/*.rb"
# orglob:  pattern: "legacy/**/*.rb"  encoding: EUC-JP  exclude: ["**/test/**", "**/spec/**"]
# orglob:  - "**/test/**/*.rb"  - pattern: "legacy/**/*.rb"    encoding: EUC-JP    exclude: ["**/test/**", "**/spec/**"]

The hash can have an optional encoding attribute. You can specify the encoding of the file by the names defined for Ruby. The list of all available encoding names can be found by the command:

ruby -e "puts Encoding.name_list"

The default value is UTF-8.

Also, the hash can have an optional exclude attribute. You can exclude any files from the pattern matched ones by this attribute.

If you write a string as a glob, the string value can be the pattern of the glob, without encoding attribute.

If you omit the glob attribute in a rule, the rule will be applied to all files given to Goodcheck.

If both your rule and its pattern has glob, Goodcheck will scan the pattern with files matching the glob condition in the pattern.

rules:  - id: glob_test    pattern:      - literal: 123 # This pattern applies to .css files        glob: "*.css"      - literal: abc # This pattern applies to .txt files    glob: "*.txt"

A rule with negated pattern#

Goodcheck rules are usually to detect something is included in a file. You can define the negated rules for the opposite, something is missing in a file.

rules:  - id: negated    not:      pattern: "<!DOCTYPE html>"    message: Write a doctype on HTML files.    glob: "**/*.html"

A rule without pattern#

You can define a rule without pattern. The rule emits an issue on each file specified with glob. You cannot omit glob from a rule definition without pattern.

rules:  - id: without_pattern    message: |      Read the operation manual for DB migration:    glob: db/schema.rb

The output of goodcheck check will be something like:

db/schema.rb:-:# This file is auto-generated from the current state of the database. Instead: Read the operation manual for DB migration:


The version 2.0.0 introduces a new abstraction to define patterns, called trigger. You can continue using a pattern in a rule, but using a trigger allows more flexible pattern definitions and more precise testing.

rules:  - id: trigger    message: Using trigger    trigger:      - pattern: <blink        glob: "**/*.html"        fail:          - <blink></blink>      - not:          pattern:            token: <meta charset="UTF-8">            case_sensitive: false        glob: "**/*.html"        pass: |          <html>            <meta charset="utf-8"></meta>          </html>

You can keep pattern definitions, but using goodcheck test against pattern with glob does not work. If your pattern definition includes glob, switching to trigger would make sense.

Importing rules#

goodcheck.yml can have an optional import attribute.

import:  - lib/goodcheck/rules.yml  - lib/goodcheck/rules/**/*.yml  - file:///usr/share/goodcheck/rules.yml  -  -

The value of import can be an array of:

  • a string or glob pattern which represents a relative file path from the config file such as goodcheck.yml
  • a file/http/https URL which represents the location of rules
  • a file/http/https URL, with a .tar.gz extension, which includes rule files

The rules file to be imported should be a YAML file with an array of rules, for example:

- id: imported_rule_1  message: Rule 1  pattern: rule-1
- id: imported_rule_2  message: Rule 2  pattern: rule-2# more rules...

Downloaded rules#

Downloaded rules are cached in the cache directory in the Goodcheck home directory. The Goodcheck home directory is ~/.goodcheck, but you can customize the location with GOODCHECK_HOME environment variable.

The cache expires in 3 minutes.

Excluding files#

goodcheck.yml can have an optional exclude or exclude_binary attribute, which allows you to exclude any files.

exclude:  - node_modules  - vendor  - assets/**/*.png
exclude_binary: true
  • exclude - allows one or more strings, representing an excluded directory or a glob pattern for excluded files.
  • exclude_binary - allows a boolean. Defaults to false. If enabled, Goodcheck will exclude files considered as binary. For example, files like foo.png or are considered as binary.


A severity represents an importance level of a rule. You can optionally specify any severity to a rule, such as error, warning, or info.

Also, you can allow only specific severities via the top-level severity option. See below:

rules:  # No problems.  - id: an-error-rule    pattern: foo    message: An error message.    severity: error
  # Goodcheck will deny this rule with `severity: info` by the `severity.allow` option below.  - id: an-info-rule    pattern: foo    message: An information message.    severity: info
  # Goodcheck will deny this rule without `severity` by the `severity.required` option below.  - id: a-rule-without-severity    pattern: foo    message: A message without a severity.
severity:  allow: [error, warning]  required: true

In summary:

  • rules[].severity - a string that represents a rule’s severity
  • severity.allow - a list of allowed severities (defaults to all allowed)
  • severity.required - a boolean value whether or not to require severities (defaults to false)

Disabling rules with inline comments#

You can disable rule warnings on a specific line using inline comments supported by common languages.

  • goodcheck-disable-line
  • goodcheck-disable-next-line

For example, for Ruby:

# goodcheck-disable-next-lineputs "Github"puts "Github" # goodcheck-disable-line

For JavaScript:

// goodcheck-disable-next-lineconsole.log("Github");console.log("Github"); // goodcheck-disable-line