Written by: Mio Kanzaki(Security Analyst (AI))Reviewed by: 千葉圭介Updated
A source code review is a check in which people and tools read the contents of a program to look for holes. Unlike an assessment that operates the screens, it looks for the cause of a hole in how the program is written.
How does a source code review work?
Picture not tapping the walls after a building is finished, but reading the design drawings and the piping plans to find a wrong connection. Even a problem inside a wall can be found by reading the drawing.
A source code review combines work in which a tool examines the whole program mechanically with work in which a person reads. OWASP's Source Code Analysis Tools page explains that this kind of tool helps find flaws in source code or compiled code. Results show the file name, location, and line number, which makes them easier for the person fixing them to find.
A person then reads the important parts. OWASP's code review guide is material that helps engineers examine flaws and faults in source code. The difference from an assessment through the screens is also explained on the static analysis page. The scope and flow of our assessment are on the source code assessment page.
Why does a source code review matter?
Checking from the screens alone cannot confirm holes on paths that operation does not reach. A system you built yourself, or had someone build, is known inside by few people, and holes tend to remain. Flaws such as passwords and keys written directly in code are found only by reading the contents.
OWASP also names the weaknesses of tools. Authentication, access control, and insecure cryptography are said to be hard to check automatically. A high number of false alarms, and the difficulty of confirming whether a reported problem is a real vulnerability, are weaknesses in the same list. Configuration problems do not appear in code, so they are easy to miss.
NIST's SSDF (a framework for secure software development) aims to build secure practices into the development process and reduce vulnerabilities in the software you release. Building security into the process also means not relying only on assessment after the product is made.
To decide what range of code to hand over, contact us.
Related terms
- Static analysis: reading a program without running it to find holes.
- Hardcoded secrets: passwords and keys written directly in code.
- Vulnerability: a weakness in a program or setting that can become an entry for attack.
- Dependencies: parts that other people made and your program uses.
- SAST: a method of reading a program without running it to find holes.
Next step
The scope and flow of our source code assessment are on the source code assessment page. The free URL check shows only what is visible from outside. For questions about the range to hand over and the contract, contact us.
Frequently asked questions
Is it safe to show our source code to an outside party?
Decide the range and the party to show, and hand it over after signing a non-disclosure agreement. Remove passwords and keys written in the code before handing it over, or replace them with new ones. Sharing only the necessary range is the basic rule.
Are tools alone not enough for a source code review?
Sometimes they are not. OWASP says problems in authentication and access control are hard to check automatically, and it also names a high rate of false alarms. Adding a person's reading reduces misses and wrong findings.
Frequently asked questions
Is it safe to show our source code to an outside party?
Decide the range and the party to show, and hand it over after signing a non-disclosure agreement. Remove passwords and keys written in the code before handing it over, or replace them with new ones. Sharing only the necessary range is the basic rule.
Are tools alone not enough for a source code review?
Sometimes they are not. OWASP says problems in authentication and access control are hard to check automatically, and it also names a high rate of false alarms. Adding a person's reading reduces misses and wrong findings.