CodeQL

CodeQL #

CodeQL is a powerful static analysis framework that allows developers and security researchers to query a codebase for specific code patterns. The CodeQL standard libraries are included with the installation and implement support for both inter- and intraprocedural control flow and data flow analysis. However, be aware that the learning curve for writing your own custom queries is steep, and documentation for the CodeQL standard libraries is still scant.

If you are planning to run CodeQL on a closed-source repository, you may need a GitHub Enterprise or GitHub Advanced Security license. (For details, see the CodeQL installation instructions and the CodeQL license.)

Benefits of using CodeQL #

  • Supports interprocedural control flow and data flow queries across the entire codebase
  • Allows for fine-grained control over the abstract syntax tree, control flow graph, and data flow graph
  • Comes with a large set of libraries and predefined queries for each supported language
  • Prevents the introduction of known bugs and security vulnerabilities into the codebase
  • Easily added to CI/CD pipelines

Ideal use case #

The following questions can help answer whether CodeQL is the right tool to identify variants of a given bug type:

  • Do you have access to the source code and any third-party dependencies, and (for compiled languages) can you build the project?
  • Are you analyzing an open-source codebase, or is the use of CodeQL covered by a GitHub Enterprise or GitHub Advanced Security license?
  • Does CodeQL support the languages used in your project?
  • Does the bug class require either fine-grained control of the abstract syntax tree, or interprocedural control flow or data flow analysis to express?
  • Is analysis time not important? (Complex, interprocedural queries may take a long time to run.)

If the answer to any of these questions is “no,” we recommend that you start out by attempting to model the bug class using a tool like Semgrep instead.

This content is licensed under a Creative Commons Attribution 4.0 International license.