Aegis Security: AI-Era Security Scanner

About the Project

Software development processes have gained unprecedented speed with the introduction of AI-powered coding assistants like Copilot, Cursor, Claude, and ChatGPT. However, this speed has introduced a completely new threat vector to the security literature: AI-Generated Vulnerabilities.

The problem is that AI assistants produce code that looks syntactically perfect, clean, and compiles without issues. Yet, under the hood, they can leave behind structural flaws that are very easy to miss by the human eye, such as business logic errors, authorization bypasses, and skipped asynchronous operations. Traditional Security Scanning (SAST) tools fall short in catching these “logical” mistakes because they rely on word-based and legacy static analysis methods.

Aegis Security is a next-generation, AI-era security scanner built exactly to solve this modern problem. It detects behavioral and logical errors made by AI assistants in milliseconds and offers remediation suggestions—before the developer even deploys the code to production.

Detailed Features and Algorithm

Aegis is not an ordinary linter; it is a semantic security engine that analyzes the code’s Abstract Syntax Tree (AST):

• AI-Focused Threat Detection: It deeply analyzes 33 different modern vulnerabilities that traditional tools miss and AI frequently falls into—such as forgotten await statements (floating promises), fake validation bypasses, hardcoded secrets, and weak cryptographic practices.
• Deep AST Analysis with Tree-sitter: Aegis reads code not as plain text, but as a tree structure. Thanks to the Tree-sitter infrastructure, it grasps the logical flow of the code, how variables are transported, and how functions interact with each other. This results in near-zero false positives.
• Multi-Language and Framework Support: It currently operates at full capacity on TypeScript, JavaScript (Node.js, React, Next.js, etc.), and Python projects. Its architecture allows for rapid integration of Go, Rust, and Java support.
• Flawless CI/CD Integration: You can add Aegis to GitHub Actions, GitLab CI, or Bitbucket Pipelines with a single command line. It automatically scans every opened Pull Request (PR) and exports its findings in SARIF (Static Analysis Results Interchange Format) to display directly in the GitHub Security tab.
• 100% Local and Privacy-Focused: Aegis does not send your code to the cloud, use any external API, or require an internet connection. It runs entirely on your own machine, completely isolated and secure.

How Does It Differ from Traditional SAST Tools?

Tools like Semgrep, SonarQube, or ESLint look for specific keywords (regex-based) or known library errors. Aegis understands the context. For instance, if a req.user object is created in an API route but used in a database query without passing through an authorization middleware, Aegis identifies this as an incomplete “Auth Bypass” hastily written by AI. It doesn’t just say “there’s an error”; it explains which “shortcut” thinking of the AI caused it.

Target Audience and Use Cases

• DevOps and DevSecOps Engineers: Those looking to add an extremely fast-running security layer to continuous integration (CI) pipelines without cluttering them with false alerts.
• Development Teams Using AI Assistants: Modern software teams that boost their coding speed using Cursor or Copilot but are wary of the security risks this speed brings.
• Open Source Maintainers: Project managers who want to quickly and securely review Pull Requests coming from the community, which are highly likely to be written with AI assistance.

Technological Infrastructure and Performance

Aegis is written entirely from scratch using the Rust programming language for maximum performance and memory safety.

• Speed: It can scan an average of 12,000 to 15,000 lines of code per second. Scanning even large monorepos takes only a few seconds.
• Deployment: It can be installed in seconds via Cargo, npm, or as a standalone binary executable. It produces standards-compliant output in JSON, plain text (CLI), and SARIF formats.

GitHub Repository: github.com/vahapogut/Aegis-Security