Module code_security 

Code security analysis for LLM-generated code outputs.

Scans LLM response text for code blocks (Markdown fenced, indented, or inline) and analyses them for common security vulnerabilities:

• SQL Injection — string concatenation in SQL queries
• Command Injection — os.system(), eval(), child_process.exec()
• Path Traversal — ../ in file operations without sanitisation
• Hardcoded Credentials — password = "...", AWS keys in code
• Insecure Deserialization — pickle.loads(), yaml.load() without SafeLoader
• XSS Patterns — innerHTML, document.write(), dangerouslySetInnerHTML
• Insecure Crypto — MD5/SHA1 for passwords, Math.random() for security

Structs

CodeSecurityAnalyzer

Analyser that scans text for code blocks and checks them for security vulnerabilities.

Enums

CodeLanguage

Programming language detected in a code block.