In this video, GitHub invites viewers to tackle AI security challenges in the Secure Code Game Season 3. Hosted by the GitHub team, it encourages ethical hacking against language models and secure code remediation.

AI Security Challenges in GitHub’s Secure Code Game Season 3

GitHub presents the third season of the Secure Code Game, a competitive platform designed for developers interested in learning about AI security threats and defensive coding techniques.

Game Overview

Objective: Outsmart a large language model (LLM) by crafting prompts that exploit vulnerabilities and cause it to leak secrets.
Remediation: After successfully attacking the LLM, participants must identify and fix the code to address the security issue, demonstrating defense-in-depth strategies.
Technology: The game uses GitHub Codespaces, making it accessible and easy to join for anyone interested in AI, cybersecurity, or secure development practices.

Key Features

AI Security Focus: Challenges center around prompt injection, vulnerability analysis, and securing LLM-based applications.
Hands-On Learning: Participants practice both offensive (ethical hacking) and defensive (code patching) skills.
Quick Sessions: Complete challenges in under a minute, making the game approachable for busy developers.
Free Participation: The Secure Code Game is open to the public at gh.io/secure-code-game.

Stay Connected

GitHub encourages viewers to stay up to date by connecting on various platforms:

About GitHub

GitHub is a global platform for software development, collaboration, and secure code sharing, serving over 100 million developers around the world.

To play or learn more, visit Secure Code Game.