Man1 - npm-audit.1

Table of Contents

NAME

npm-audit - Run a security audit

Synopsis

    npm audit [--json] [--production] [--audit-level=(low|moderate|high|critical)]
    npm audit fix [--force|--package-lock-only|--dry-run|--production|--only=(dev|prod)]

    common options: [--production] [--only=(dev|prod)]

Description

The audit command submits a description of the dependencies configured in your project to your default registry and asks for a report of known vulnerabilities. If any vulnerabilities are found, then the impact and appropriate remediation will be calculated. If the fix argument is provided, then remediations will be applied to the package tree.

The command will exit with a 0 exit code if no vulnerabilities were found.

Note that some vulnerabilities cannot be fixed automatically and will require manual intervention or review. Also note that since npm audit fix runs a full-fledged npm install under the hood, all configs that apply to the installer will also apply to npm install – so things like npm audit fix –package-lock-only will work as expected.

By default, the audit command will exit with a non-zero code if any vulnerability is found. It may be useful in CI environments to include the –audit-level parameter to specify the minimum vulnerability level that will cause the command to fail. This option does not filter the report output, it simply changes the command’s failure threshold.

Audit Endpoints

There are two audit endpoints that npm may use to fetch vulnerability information: the Bulk Advisory endpoint and the Quick Audit endpoint.

Bulk Advisory Endpoint

As of version 7, npm uses the much faster Bulk Advisory endpoint to optimize the speed of calculating audit results.

npm will generate a JSON payload with the name and list of versions of each package in the tree, and POST it to the default configured registry at the path /-/npm/v1/security/advisories/bulk .

Any packages in the tree that do not have a version field in their package.json file will be ignored. If any –omit options are specified (either via the –omit config, or one of the shorthands such as –production, –only=dev, and so on), then packages will be omitted from the submitted payload as appropriate.

If the registry responds with an error, or with an invalid response, then npm will attempt to load advisory data from the Quick Audit endpoint.

The expected result will contain a set of advisory objects for each dependency that matches the advisory range. Each advisory object contains a name, url, id, severity, vulnerable_versions, and title .

npm then uses these advisory objects to calculate vulnerabilities and meta-vulnerabilities of the dependencies within the tree.

Quick Audit Endpoint

If the Bulk Advisory endpoint returns an error, or invalid data, npm will attempt to load advisory data from the Quick Audit endpoint, which is considerably slower in most cases.

The full package tree as found in package-lock.json is submitted, along with the following pieces of additional metadata:

  • npm_version
  • node_version
  • platform
  • arch
  • node_env

All packages in the tree are submitted to the Quick Audit endpoint. Omitted dependency types are skipped when generating the report.

Scrubbing

Out of an abundance of caution, npm versions 5 and 6 would “scrub” any packages from the submitted report if their name contained a / character, so as to avoid leaking the names of potentially private packages or git URLs.

However, in practice, this resulted in audits often failing to properly detect meta-vulnerabilities, because the tree would appear to be invalid due to missing dependencies, and prevented the detection of vulnerabilities in package trees that used git dependencies or private modules.

This scrubbing has been removed from npm as of version 7.

Calculating Meta-Vulnerabilities and Remediations

npm uses the @npmcli/metavuln-calculator http://npm.im/@npmcli/metavuln-calculator module to turn a set of security advisories into a set of “vulnerability” objects. A “meta-vulnerability” is a dependency that is vulnerable by virtue of dependence on vulnerable versions of a vulnerable package.

For example, if the package foo is vulnerable in the range >=1.0.2 <2.0.0, and the package bar depends on foo@^1.1.0, then that version of bar can only be installed by installing a vulnerable version of foo . In this case, bar is a “metavulnerability”.

Once metavulnerabilities for a given package are calculated, they are cached in the ~/.npm folder and only re-evaluated if the advisory range changes, or a new version of the package is published (in which case, the new version is checked for metavulnerable status as well).

If the chain of metavulnerabilities extends all the way to the root project, and it cannot be updated without changing its dependency ranges, then npm audit fix will require the –force option to apply the remediation. If remediations do not require changes to the dependency ranges, then all vulnerable packages will be updated to a version that does not have an advisory or metavulnerability posted against it.

Exit Code

The npm audit command will exit with a 0 exit code if no vulnerabilities were found. The npm audit fix command will exit with 0 exit code if no vulnerabilities are found or if the remediation is able to successfully fix all vulnerabilities.

If vulnerabilities were found the exit code will depend on the audit-level configuration setting.

Examples

Scan your project for vulnerabilities and automatically install any compatible updates to vulnerable dependencies:

    $ npm audit fix

Run audit fix without modifying node_modules, but still updating the pkglock:

    $ npm audit fix --package-lock-only

Skip updating devDependencies:

    $ npm audit fix --only=prod

Have audit fix install SemVer-major updates to toplevel dependencies, not just SemVer-compatible ones:

    $ npm audit fix --force

Do a dry run to get an idea of what audit fix will do, and also output install information in JSON format:

    $ npm audit fix --dry-run --json

Scan your project for vulnerabilities and just show the details, without fixing anything:

    $ npm audit

Get the detailed audit report in JSON format:

    $ npm audit --json

Fail an audit only if the results include a vulnerability with a level of moderate or higher:

    $ npm audit --audit-level=moderate

Configuration

<!– AUTOGENERATED CONFIG DESCRIPTIONS START –> <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

audit-level

  • Default: null
  • Type: null, “info”, “low”, “moderate”, “high”, “critical”, or “none”

The minimum level of vulnerability for npm audit to exit with a non-zero exit code. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

dry-run

  • Default: false
  • Type: Boolean

Indicates that you don’t want npm to make any changes and that it should only report what it would have done. This can be passed into any of the commands that modify your local installation, eg, install, update, dedupe, uninstall, as well as pack and publish .

Note: This is NOT honored by other network related commands, eg dist-tags, owner, etc. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

force

  • Default: false
  • Type: Boolean

Removes various protections against unfortunate side effects, common mistakes, unnecessary performance degradation, and malicious input.

  • Allow clobbering non-npm files in global installs.
  • Allow the npm version command to work on an unclean git repository.
  • Allow deleting the cache folder with npm cache clean .
  • Allow installing packages that have an engines declaration requiring a different version of npm.
  • Allow installing packages that have an engines declaration requiring a different version of node, even if –engine-strict is enabled.
  • Allow npm audit fix to install modules outside your stated dependency range (including SemVer-major changes).
  • Allow unpublishing all versions of a published package.
  • Allow conflicting peerDependencies to be installed in the root project.
  • Implicitly set –yes during npm init .
  • Allow clobbering existing values in npm pkg

If you don’t have a clear idea of what you want to do, it is strongly recommended that you do not use this option! <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

json

  • Default: false
  • Type: Boolean

Whether or not to output JSON data, rather than the normal output.

  • In npm pkg set it enables parsing set values with JSON.parse() before saving them to your package.json .

Not supported by all npm commands. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

package-lock-only

  • Default: false
  • Type: Boolean

If set to true, the current operation will only use the package-lock.json, ignoring node_modules .

For update this means only the package-lock.json will be updated, instead of checking node_modules and downloading dependencies.

For list this means the output will be based on the tree described by the package-lock.json, rather than the contents of node_modules . <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

omit

  • Default: ’dev’ if the NODE_ENV environment variable is set to  ’production’, otherwise empty.
  • Type: “dev”, “optional”, or “peer” (can be set multiple times)

Dependency types to omit from the installation tree on disk.

Note that these dependencies are still resolved and added to the package-lock.json or npm-shrinkwrap.json file. They are just not physically installed on disk.

If a package type appears in both the –include and –omit lists, then it will be included.

If the resulting omit list includes * ’dev’, then the *NODE_ENV environment variable will be set to * ’production’* for all lifecycle scripts. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

workspace

  • Default:
  • Type: String (can be set multiple times)

Enable running a command in the context of the configured workspaces of the current project while filtering by running only the workspaces defined by this configuration option.

Valid values for the workspace config are either:

  • Workspace names
  • Path to a workspace directory
  • Path to a parent workspace directory (will result in selecting all workspaces within that folder)

When set for the npm init command, this may be set to the folder of a workspace which does not yet exist, to create the folder and set it up as a brand new workspace within the project.

This value is not exported to the environment for child processes. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

workspaces

  • Default: null
  • Type: null or Boolean

Set to true to run the command in the context of all configured workspaces.

Explicitly setting this to false will cause commands like install to ignore workspaces altogether. When not set explicitly:

  • Commands that operate on the node_modules tree (install, update, etc.) will link workspaces into the node_modules folder. - Commands that do other things (test, exec, publish, etc.) will operate on the root project, unless one or more workspaces are specified in the workspace config.

This value is not exported to the environment for child processes. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

include-workspace-root

  • Default: false
  • Type: Boolean

Include the workspace root when workspaces are enabled for a command.

When false, specifying individual workspaces via the workspace config, or all workspaces via the workspaces flag, will cause npm to operate only on the specified workspaces, and not on the root project. <!– automatically generated, do not edit manually –> <!– see lib/utils/config/definitions.js –>

<!– AUTOGENERATED CONFIG DESCRIPTIONS END –>

See Also

  • npm help install
  • npm help config

Author: dt

Created: 2022-02-22 Tue 16:08