security

Subscribe to all “security” posts via RSS or follow GitHub Changelog on Twitter to stay updated on everything we ship.

~ cd github-changelog
~/github-changelog|main git log main
showing all changes successfully

In February 2022, we introduced experimental CodeQL queries that utilize machine learning to identify more potential vulnerabilities. This feature was only available for JavaScript / TypeScript code and was available to code scanning users that enabled the optional security-extended or security-and-quality query suites.

We disabled this experimental feature for new code scanning users in June 2023. Today, we're sunsetting it for all users.

Any currently open code scanning alerts from these queries (Rule ID starts with js/ml-powered/) will be closed. Closed alerts will still be visible in the code scanning alerts view in your repository’s Security tab. The complete history of each alert will remain accessible by clicking on the alert.

CodeQL will continue to run the existing non-ML versions of these queries and provide you with highly precise and actionable alerts.

We’ve learned a lot from the feedback and experience of the repositories that participated in this experiment, and we’ve since ramped up our investment in AI-powered security technology. This new technology is already boosting our ability to cover more sources and sinks of untrusted data in order to significantly increase the coverage and depth of all queries.

See more

On September 27, 2023, we began blocking npm package publishes with differing name or version fields between the manifest and tarball package.json. This blocking protects against obfuscation. The different fields in the manifests have been assessed from a risk-based perspective. We will continue to analyze for other mismatches that can be blocked that won’t have adverse effects on the ecosystem. If a package is blocked, a user may receive an error message similar to “Package ‘version’ is “1.0.4”. It should match “1.0.3” from “package.json” in packaged tarball. Make all changes to package.json before packaging a tarball to publish.” In addition, a new tool, npm pkg fix, can help users fix any validation errors from the registry when they attempt to publish a package.

See more

npm provenance is now generally available.

npm packages built on a supported cloud CI/CD system can publish with provenance. Today this includes GitHub Actions and GitLab CI/CD.

Publishing with provenance verifiably links the package back to its source repository and build instructions. Provenance is restricted to public packages and public source repositories only.

npm will check the linked source commit and repository when you view a package's provenance information on npmjs.com. If the linked source commit or repository cannot be found, an error displays at the top of the page and alongside the provenance information to let you know that provenance for this package can no longer be established. This can happen when a repository is deleted or made private.

Once published, packages display provenance on the registry website:

Provenance displayed on the registry website

For more information, see generating provenance.

See more

Passkeys are a replacement for passwords when signing in, providing higher security, ease-of-use, and loss-protection. They are now generally available on GitHub.com for all users. By using a passkey you no longer need to enter a password, or even your username, when you sign in – nor do you need to perform 2FA, if you have 2FA enabled on your account. This is because passkeys validate your identity, as well as possession of a device, so they count as two authentication factors in one. Once enrolled, you can register a brand new passkey and upgrade many security keys to passkeys.

Screenshot of the security key upgrade prompt, asking the user if they'd like to upgrade a security key called 'fingerprint' to a passkey.

To learn more, check out our documentation "About passkeys", as well as this previous blog post from the passkeys beta announcement. If you have any feedback, please drop us a note in our public discussion – we're excited for this advance in account security, and would love to understand how we can make it better for you.

See more

Code scanning with CodeQL now supports Java codebases that use Project Lombok. Previously, code scanning users were able to scan Java applications that contained Lombok code, but all the contents of files containing Lombok code were either skipped or users had to apply a workaround to prepare the applications for scanning. The improved support means that code with Lombok features will be automatically scanned without requiring any workaround.

As more code will now be analyzed by the CodeQL engine, we can establish more accurate data flow (or lack thereof) through Lombok code. This might have an impact on the number of alerts produced by a scan. The most common scenario is that additional alerts appear in the newly-analyzed code. Conversely, there is a very small chance that some existing alerts are closed.

Improved support for Java applications built using Lombok is available for code scanning users on GitHub.com starting today and GitHub Enterprise Server users starting with 3.11. CodeQL CLI will provide out of the box support starting with the upcoming version 2.14.4. Security researchers can set up the CodeQL CLI and VS Code extension by following these instructions.

See more

We have released a new API for people who write custom CodeQL queries which make use of dataflow analysis. The new API offers additional flexibility, improvements that prevent common pitfalls with the old API, and improves query evaluation performance by 5%. Whether you’re writing CodeQL queries for personal interest, or are participating in the bounty programme to help us secure the world’s code: this post will help you move from the old API to the new one.

This API change is relevant only for users who write their own custom CodeQL queries. Code scanning users who use GitHub’s standard CodeQL query suites will not need to make any changes.

With the introduction of the new dataflow API, the old API will be deprecated. The old API will continue to work until December 2024; the CodeQL CLI will start emitting deprecation warnings in December 2023.

To demonstrate how to update CodeQL queries from the old to the new API, consider this example query which uses the soon-to-be-deprecated API:

class SensitiveLoggerConfiguration extends TaintTracking::Configuration {
  SensitiveLoggerConfiguration() { this = "SensitiveLoggerConfiguration" } // 6: characteristic predicate with dummy string value (see below)

  override predicate isSource(DataFlow::Node source) { source.asExpr() instanceof CredentialExpr }

  override predicate isSink(DataFlow::Node sink) { sinkNode(sink, "log-injection") }

  override predicate isSanitizer(DataFlow::Node sanitizer) {
    sanitizer.asExpr() instanceof LiveLiteral or
    sanitizer.getType() instanceof PrimitiveType or
    sanitizer.getType() instanceof BoxedType or
    sanitizer.getType() instanceof NumberType or
    sanitizer.getType() instanceof TypeType
  }

  override predicate isSanitizerIn(DataFlow::Node node) { this.isSource(node) }
}

import DataFlow::PathGraph

from SensitiveLoggerConfiguration cfg, DataFlow::PathNode source, DataFlow::PathNode sink
where cfg.hasFlowPath(source, sink)
select sink.getNode(), source, sink, "This $@ is written to a log file.",
 source.getNode(),
  "potentially sensitive information"

To convert the query to the new API:

  1. You use a module instead of a class. A CodeQL module does not extend anything, it instead implements a signature. For both data flow and taint tracking configurations this is DataFlow::ConfigSig or DataFlow::StateConfigSigif FlowState is needed.
  2. Previously, you would choose between data flow or taint tracking by extending DataFlow::Configuration or TaintTracking::Configuration. Instead, now you define your data or taint flow by instantiating either the DataFlow::Global<..> or TaintTracking::Global<..> parameterized modules with your implementation of the shared signature and this is where the choice between data flow and taint tracking is made.
  3. Predicates no longer override anything, because you are defining a module.
  4. The concepts of sanitizers and barriers are now unified under isBarrier and it applies to both taint tracking and data flow configurations. You must use isBarrier instead of isSanitizer and isBarrierIn instead of isSanitizerIn.
  5. Similarly, instead of the taint tracking predicate isAdditionalTaintStep you use isAdditionalFlowStep .
  6. A characteristic predicate with a dummy string value is no longer needed.
  7. Do not use the generic DataFlow::PathGraph. Instead, the PathGraph will be imported directly from the module you are using. For example, SensitiveLoggerFlow::PathGraph in the updated version of the example query below.
  8. Similar to the above, you’ll use the PathNode type from the resulting module and not from DataFlow.
  9. Since you no longer have a configuration class, you’ll use the module directly in the from and where clauses. Instead of using e.g. cfg.hasFlowPath or cfg.hasFlow from a configuration object cfg, you’ll use flowPath or flow from the module you’re working with.

Taking all of the above changes into account, here’s what the updated query looks like:

module SensitiveLoggerConfig implements DataFlow::ConfigSig {  // 1: module always implements DataFlow::ConfigSig or DataFlow::StateConfigSig
  predicate isSource(DataFlow::Node source) { source.asExpr() instanceof CredentialExpr } // 3: no need to specify 'override'
  predicate isSink(DataFlow::Node sink) { sinkNode(sink, "log-injection") }

  predicate isBarrier(DataFlow::Node sanitizer) {  // 4: 'isBarrier' replaces 'isSanitizer'
    sanitizer.asExpr() instanceof LiveLiteral or
    sanitizer.getType() instanceof PrimitiveType or
    sanitizer.getType() instanceof BoxedType or
    sanitizer.getType() instanceof NumberType or
    sanitizer.getType() instanceof TypeType
  }

  predicate isBarrierIn(DataFlow::Node node) { isSource(node) } // 4: isBarrierIn instead of isSanitizerIn

}

module SensitiveLoggerFlow = TaintTracking::Global<SensitiveLoggerConfig>; // 2: TaintTracking selected 

import SensitiveLoggerFlow::PathGraph  // 7: the PathGraph specific to the module you are using

from SensitiveLoggerFlow::PathNode source, SensitiveLoggerFlow::PathNode sink  // 8 & 9: using the module directly
where SensitiveLoggerFlow::flowPath(source, sink)  // 9: using the flowPath from the module 
select sink.getNode(), source, sink, "This $@ is written to a log file.", source.getNode(),
  "potentially sensitive information"

While not covered in this example, you can also implement the DataFlow::StateConfigSig signature if flow-state is needed. You then instantiate DataFlow::GlobalWithState or TaintTracking::GlobalWithState with your implementation of that signature. Another change specific to flow-state is that instead of using DataFlow::FlowState, you now define a FlowState class as a member of the module. This is useful for using types other than string as the state (e.g. integers, booleans). An example of this implementation can be found here.

This functionality is available with CodeQL version 2.13.0. If you would like to get started with writing your own custom CodeQL queries, follow these instructions to get started with the CodeQL CLI and the VS Code extension.

See more

You can now use the REST API to get global security advisories from the Advisory Database. This makes it easy to get access to the Advisory Database's free, open source list of actionable security advisories and CVEs which include machine readable mappings to the ecosystem, package name, and affected versions of impacted software.

Learn more about GitHub's global security advisories and the Advisory Database.

See more

CodeQL is the analysis engine that powers GitHub code scanning for over 100,000 repositories. We continuously improve our analysis capabilities, language support and performance to help open source developers and enterprises catch vulnerabilities before they make their way into production code. CodeQL is also an instrumental tool for the security researcher community and was used to identify 36 new CVE.

We release updates and improvements for CodeQL on a regular basis. We don’t get to call out all the improvements, but we want to highlight some of the most important updates we’ve shipped for CodeQL in the first half of the year:

  • Shortly before WWDC in June, we added beta support for Swift, which together with Kotlin completes CodeQL’s support for next-generation mobile development.
  • We’ve updated CodeQL to support these new language versions (view all): Swift 5.8.1, C#11 , .NET 7, Kotlin 1.8, Go 1.20, TypeScript 5.0 & 5.1, Ruby 3.2, Java 20.
  • We saw a 16% average performance improvement for CodeQL analyses.
  • We improved CodeQL modelling for popular Ruby libraries (SQLite, MySQL, Rack) and added coverage for more than 5000 API methods in Java, increasing analysis coverage and reducing false negatives.
  • We released a new mechanism called default setup, to configure CodeQL at the repo and the organization level.
  • We added 4 new memory-corruption queries for C/C++, 6 new queries for Java, 1 for Python and adjusted over 100 queries across all languages.
  • We started showing actionable information on the tool status page.
  • Enabled scanning Python repositories without installing dependencies.
  • We made the release process faster, 1 week, and optimised the roll-out strategy to get you on the latest release as quick as possible, benefiting from the latest updates in CodeQL.
  • Deprecated CodeQL Action v1 and enabled Dependabot to automatically move you to a newer version.

These features have been shipped across multiple versions of CodeQL from 2.12.0 up to 2.14.0, which are shipped with GHES 3.9 and upcoming 3.10. All users of CodeQL code scanning on GitHub.com automatically benefit from the latest improvements.

See more

Starting today, publishing with provenance is restricted to public source repositories only. Private source repositories are no longer supported for use with provenance for public packages.

As announced on July 11, 2023: npm will verify the linked source commit and repository when users view a package's provenance information on npmjs.com. If the linked source commit or repository cannot be found, an error will be displayed. This can occur if a repository is deleted or if it is made private.

Read more about viewing npm provenance and publishing with provenance.

See more

Passkeys are a replacement for passwords when signing in, providing higher security, ease-of-use, and loss-protection. They're now available on GitHub.com as a public beta – see this blog post for more information.

This public beta is open to all users with a password, regardless of whether you use 2FA. To get started, enable passkeys as a feature preview.

By using passkeys, you no longer need to enter a password, or even your username, when you sign in – nor do you need to perform 2FA, if you have 2FA enabled on your account. That's because passkeys validate your identity, as well as possession of a device, so they count as two authentication factors in one.

Once enrolled, you can register a brand new passkey and upgrade many security keys to a passkey. If you're enrolled in the preview, the next time you use an eligible security key you'll be asked to upgrade it.
Screenshot of the security key upgrade prompt, asking the user if they'd like to upgrade a security key called 'fingerprint' to a passkey.

To learn more, check out this blog post about passkeys, as well as "About passkeys" in our documentation. If you have any feedback, please drop us a note in our public discussion – we're excited for this advance in account security, and would love to understand how we can make it better for you.

See more