This site is the archived OWASP Foundation Wiki and is no longer accepting Account Requests.
To view the new OWASP Foundation website, please visit

Automated Code Review

Revision as of 11:01, 27 July 2006 by EoinKeary (talk | contribs) (Preface)

Jump to: navigation, search
OWASP Code Review Guide Table of Contents


While manual code reviews can find security flaws in code, they suffer from two problems. Manual code reviews are slow, covering 100-200 lines per hour on average. Also, there are hundreds of security flaws to look for in code, while humans can only keep about seven items in memory at once. Source code analysis tools can search a program for hundreds of different security flaws at once at a rate far greater than any human can review code. However, these tools don't eliminate the need for a human reviewer, as they produce both false positive and false negative results.

Reasons for using automated tools:

In large scale code review operations for enterprises such that the volume of code is enormous automated code review techniques can assist in improving the throughput of the code review process.

Education and cultural change:

Educating developers to write secure code is the paramount goal of a secure code review. Taking code review from this standpoint is the only way to promote and improve code quality. Part of the education process is to empower devlopers with the knowledge in order to write better code.

This can be done by providing developers with a controlled set of rules which the developer can compare their code to. Automated tools provide this functionality and also help reducing the overhead from a time perspective. A developer can check his/her code using a tool without much initial knowledge of the security concerns pertaining to their task at hand. Also running a tool to assess the code if a fairly painless task once the developer becomes familiar wth the tool(s).

Tool Deployment model:

Deploying code review tools to developers helps the throughput of a code review team by helping to identify and hopefully remove most of the common and simple coding mistakes prior to a security consultant viewing the code.

This methodology improves developer knowledge and also the security consultant can spend time looking for more abstract vulerabilities.