About coding guidelines

You can customize Copilot code review with custom coding guidelines written in natural language. For more information on Copilot code review, seeUsing GitHub Copilot code review.

With coding guidelines, Copilot can give feedback based on your organization's specific coding style and best practices.

Because Copilot code review is powered by a large language model, it can help with enforcing coding guidelines that are not covered by your linter or static analysis tool.

Coding guidelines are configured at the repository level. You can create and enable up to 6 coding guidelines per repository.

Dos and don'ts for coding guidelines

• Do use simple, clear and concise language to describe your coding guideline.
• Do be as specific as possible about what Copilot should look for - that is, what you do or don't want to see in your code.
• Do take a look at theCoding guidelines examples below for some inspiration.
• Don't try to use coding guidelines to enforce style guidelines that can be covered by your linter or static analysis tool.
• Don't use wording that is ambiguous or could be interpreted in different ways.
• Don't try to fit multiple different ideas into a single coding guideline.

Creating a coding guideline

1. On GitHub, navigate to the main page of the repository.

2. Under your repository name, click Settings. If you cannot see the "Settings" tab, select the dropdown menu, then clickSettings.

   

3. In the "Code & automation" section of the sidebar, click Copilot, thenCode review.

4. ClickCreate guideline.

5. Under "Name," give the coding guideline a name.

6. Under "Description," provide a description of the coding guideline up to 600 characters long. This will be used by Copilot to understand your coding style and to decide when to leave a comment.

   How you write your description has a big impact on the quality of comments that Copilot will generate. For help with writing effective coding guidelines, seeDos and don'ts for coding guidelines above, andCoding guidelines examples below.

7. Optionally, limit the coding guideline to specific file types or paths by clickingAdd file path and adding path patterns.

   You can usefnmatch syntax to define paths to target, with* as a wildcard to match any string of characters.

   Because GitHub uses theFile::FNM_PATHNAME flag for theFile.fnmatch syntax, the* wildcard does not match directory separators (/). For example,qa/* will match all branches beginning withqa/ and containing a single slash, but will not matchqa/foo/bar. You can include any number of slashes afterqa withqa/**/*, which would match, for example,qa/foo/bar/foobar/hello-world. You can also extend theqa string withqa**/**/* to make the rule more inclusive.

   For more information about syntax options, see thefnmatch documentation.

8. Test your coding guideline to make sure it works as expected.

   1. ClickAdd sample.
   2. Add your own sample, or press Generate code sample to automatically generate a code sample based on your title and description.
   3. ClickSave to save the code sample.
   4. Test the coding guideline against your sample by pressing Run.

9. Save your coding guideline, and turn it on, by clickingSave guideline.

Running a review with coding guidelines

When you request a review from Copilot, it will automatically use the repository's enabled coding guidelines to review your code. For more information, seeUsing GitHub Copilot code review.

Comments generated based on a coding guideline will include a message, highlighting their source.

Coding guidelines examples

Example 1: Avoid using magic numbers

Title:Avoid using magic numbers

Description:Don't use magic numbers in code. Numbers should be defined as constants or variables with meaningful names.

Path patterns:**/*.py

Example 2: Don't useSELECT * in SQL queries

Title:Don't use `SELECT *` in SQL queries

Description:Don't use `SELECT *` in SQL queries. Always specify the columns you want to select. `COUNT(*)` is allowed.

Path patterns: None (applies to all file types, as SQL queries may be embedded in code).

Example 3: Usefetch for HTTP requests

Title:Use `fetch` for HTTP requests

Description:Use `fetch` for HTTP requests, not `axios` or `superagent` or other libraries.

Path patterns:**/*.ts,**/*.js,**/*.jsx,**/*.tsx

Example 4: Always tag metrics with the current environment

Title:Always tag metrics with the current environment

Description:Always include a `env` tag with the current environment when emitting metrics, for example, `env:prod` or `env:dev`.

Path patterns:*/*.go,*/*.java