Detect complex code patterns using semantic grep, by Colleen Dai

This recording comes from our OWASP Women in AppSec, Diversity and Inclusion virtual meetup on Saturday, November 14: https://www.meetup.com/womeninappsec/events/273489892/


Abstract:

We’ll discuss a program analysis tool we’re developing called Semgrep. It's a multilingual semantic tool for writing security and correctness queries on source code (for Python, Java, Go, C, and JS) with a simple “grep-like” interface. The original author, Yoann Padioleau, worked on Semgrep’s predecessor, Coccinelle, for Linux kernel refactoring, and later developed Semgrep while at Facebook. He’s now full time with us at r2c.

Semgrep is a free open-source program analysis toolkit that finds bugs using custom analysis we’ve written and OSS code checks. Semgrep is ideal for security researchers, product security engineers, and developers who want to find complex code patterns without extensive knowledge of ASTs or advanced program analysis concepts.

For example, find subprocess calls with shell=True in Python using the query:
subprocess.open(..., shell=True)
This will even find snippets like:
import subprocess as s
s.open(f'rm {args}', shell=True)

Or find hardcoded credentials using the query:
boto3.client(..., aws_secret_access_key=”...”, aws_access_key_id=”...” )

Source code: https://github.com/returntocorp/semgrep
Test in your browser: https://semgrep.dev/

Speaker bio:

Colleen Dai is a security software engineer at r2c, a startup working on building static analysis tools that focus on precision and being custom-fit to the consumer. At r2c, Colleen has worked on the language parsing along with AST matching. She is also writing rules to find security vulnerabilities in open source code. Colleen recently received her B.S. in Computer Science and M.S. in Statistics from Stanford. She regularly enjoys Brazilian Jiujitsu, drawing, and trying (and failing) not to eat everything in her fridge.