AI Coding Is Booming.
So Are Its Risks.

Discover How to Secure the New Era of “Vibe Coding” Before It's Too Late

Whether you know it or not, your developers are using tools such as Cursor, Windsurf, Zed and GitHub CoPilot to create, enhance, and fix their code — but the resulting code as well as the mere inclusion of such tools with the IDE are creating new risks, fast.

What You’ll Learn in This Whitepaper:

  • Inside the Vibe Coding Explosion: Why AI-assisted development is transforming how software is written — and why it's now the top security blind spot.
  • Groundbreaking Research Findings: Backslash tested code LLMs from OpenAI, Google, and Anthropic. The verdict? Without specific security guardrails, LLMs generated insecure code up to 90% of the time.
  • The Real Risks: Outdated AI training data, unvetted MCP servers, hallucinated vulnerabilities, LLM poisoning, and third-party flaws embedded directly into your code.
  • Addressing The Risks: How a modern approach to AppSec can preempt and mitigate the risks that traditional scanners and testing tools fail to address

‍

Get the whitepaper >
Back to Feed

Securing AI-Generated Code and Vibe Coding IDEs Whitepaper

-

May 28, 2025

What You’ll Learn in This Whitepaper:

  • Inside the Vibe Coding Explosion: Why AI-assisted development is transforming how software is written — and why it's now the top security blind spot.
  • Groundbreaking Research Findings: Backslash tested code LLMs from OpenAI, Google, and Anthropic. The verdict? Without specific security guardrails, LLMs generated insecure code up to 90% of the time.
  • The Real Risks: Outdated AI training data, unvetted MCP servers, hallucinated vulnerabilities, LLM poisoning, and third-party flaws embedded directly into your code.
  • Addressing The Risks: How a modern approach to AppSec can preempt and mitigate the risks that traditional scanners and testing tools fail to address

‍