Skip to main content

    Ignore files, folders, and code

    Exclude or skip specific files and folders from the scope of Semgrep scans in your repository or working directory. Ignoring in this context means that Semgrep does not generate findings for the ignored files and folders.

    Reference summary

    MethodUsageExamples
    To ignore blocks of code: nosemgrepCreate a comment, followed by a space ( ), followed by nosemgrep at the first line or preceding line of the pattern match. This generates a finding that is automatically ignored. // nosemgrep                   // nosemgrep: rule-id
    # nosemgrep
    To ignore files and folders: .semgrepignoreCreate a .semgrepignore file in your repository's root directory or your project's working directory and add patterns for files and folders there. Patterns follow .gitignore syntax with some caveats. See Defining ignored files and folders in .semgrepignore..semgrepignore sample file

    Understanding Semgrep defaults

    Without user customization, Semgrep refers to the following to define ignored files and folders:

    • Semgrep's default .semgrepignore file
    • Your repository's .gitignore file (if it exists)

    In the absence of a user-generated .semgrepignore, Semgrep refers to its repository's default template:

    DEFAULT_SEMGREPIGNORE_TEXT

    caution

    The default .semgrepignore file causes Semgrep to skip these folders:

    • /tests, /test
    • /vendors

    To include the above folders, create a .semgrepignore file without those paths.

    Files, folders, and code beyond Semgrep's scope

    There are files that Semgrep ignores even without .semgrepignore:

    • Large files (maximum file size defaults to 1 MB)
    • Binary files
    • Unknown file extensions (file extensions not matched with any supported programming language)

    Large files and unknown file extensions are included or excluded through command line flags (See CLI reference). Binary files are never scanned.

    This document defines files, folders and code as those that are relevant to a Semgrep scan. For example, .jpg files are not a part of Semgrep's scope and therefore are not part of the scope of this document.

    Customizing ignore behavior

    Semgrep provides several methods to customize ignore behavior. Refer to the following table to see which method suits your goal:

    GoalMethod
    To scan all files within Semgrep's scope each time you run Semgrep (only files within .git are ignored).Create an empty .semgrepignore file in your repository root directory or in your project's working directory. An empty .semgrepignore will make Semgrep scan paths in .gitignore.
    To ignore files and folders in .gitignore.Add :include .gitignore to your .semgrepignore file.
    To ignore custom files and folders each time you run a scan.Add these files to your .semgrepignore file.
    To ignore specific code blocks each time you run a scan.Create a comment with the word nosemgrep.
    To ignore files or folders for a particular scan.Run Semgrep with the flag --exclude followed by the pattern or file to be excluded. See CLI reference.
    To include files or folders for a particular scan.Run Semgrep with the flag --include followed by the pattern or file to be included. See CLI reference. When including a pattern from a .gitignore or .semgrepignore file, --include does not override either, still resulting in the file's exclusion.
    To include files or folders defined within a .gitignore for a particular scan.Run Semgrep with the flag --no-git-ignore, which overrides its definition within .semgrepignore as well.
    To ignore files or folders for a particular rule.Edit the rule to set the paths key with one or more patterns. See Rule syntax.

    Defining ignored files and folders in .semgrepignore

    .semgrepignore syntax mirrors .gitignore syntax, with the following modifications:

    • "Include" patterns (lines starting with !) are unsupported.
    • "Character range" patterns (lines including a collection of characters inside brackets) are unsupported.
    • An :include ... directive is added, which allows another file to be included in the ignore pattern list; typically this included file would be the project .gitignore. No attempt at cycle detection is made.
    • Any line that begins with a colon, but not :include, raises an error.
    • \: is added to escape leading colons.
    tip

    To ignore an entire directory, make sure the .semgrepignore entry ends with a /, designating it as a directory and not a file.

    Unsupported patterns are silently removed from the pattern list (this is done so that .gitignore files may be included without raising errors). The removal is logged.

    For a description of .gitignore syntax, see .gitignore documentation.

    Ignoring code through nosemgrep

    To ignore blocks of code, define an inline comment, followed by a space ( ), followed by the word nosemgrep at either the first line or the line preceding the potential match. Semgrep ignores all rule pattern matches. This functionality works across all supported languages.

    caution

    Ignoring code through this method still generates a finding. The finding is automatically set to the Ignored triage state.

    nosemgrep in Python:


    bad_func1() # nosemgrep

    # nosemgrep
    bad_func2()

    nosemgrep in JavaScript:


    // nosemgrep
    bad_func1()

    bad_func2(); // nosemgrep

    bad_func3( // nosemgrep
    arg
    );

    info

    The space ( ) before nosemgrep is required for Semgrep to detect this annotation.

    To ignore blocks of code for a particular rule, enter its rule-id as follows: nosemgrep: RULE_ID. To ignore multiple rules, use a comma-delimited list. rule-ids must be referenced with their namespace.

    Python examples:


    bad_func1() # nosemgrep: rule-id-1

    # nosemgrep: rule-id-1, rule-id-2
    bad_func2()

    JavaScript examples wherein rules are stored in a configs subdirectory:


    // nosemgrep: configs.rule-id-3
    bad_func1()

    bad_func2(); // nosemgrep: configs.rule-id-3

    bad_func3( // nosemgrep: configs.rule-id-3, configs.rule-id-4
    arg
    );

    info

    Previous annotations for ignoring code inline, such as nosem, are deprecated.

    Known issues

    --no-git-ignore is overridden due to default ignore patterns (.semgrepignore) (#4537)

    To fix this, create an empty .semgrepignore file. If the scan is a one-off event, delete the .semgrepignore file to restore default ignore patterns.


    Not finding what you need in this doc? Ask questions in our Community Slack group, or see Support for other ways to get help.