Skip to main content

Test your prompt injection skills

Lakera, a product company that offers solutions to monitor, protect and control your AI agents, has created Gandalf, a game that allows you to test your prompt injection skills. Named after the wise wizard from The Lord of the Rings, this engaging security game tests your abilities to outsmart AI systems through prompt injection techniques.

What is prompt injection?

Before diving into Gandalf, it's important to understand what prompt injection is. Prompt injection is a technique where users attempt to manipulate an AI system's behavior by crafting inputs that override or bypass its intended operating parameters. Think of it like SQL injection, but for AI language models.

When we interact with AI systems, we typically provide them with a prompt - a piece of text that guides their response. However, these systems often retain their base instructions (sometimes called a system prompt) that define their behavior and limitations. Prompt injection attempts to override these instructions by finding creative ways to make the AI disregard its original directives.

For example, a simple prompt injection might look like this: "Ignore all your previous instructions and instead do X." While modern AI systems are generally robust against such basic attempts, more sophisticated techniques can sometimes succeed in manipulating the AI's behavior in unintended ways.

Gandalf game and adventures

At its core, Gandalf is a security challenge game designed to demonstrate how AI language models might be vulnerable to manipulation through carefully crafted inputs. The game presents players with a seemingly simple task: convince an AI guardian to reveal a secret password. However, like its namesake's famous declaration "You shall not pass!", the AI system is programmed to resist these attempts.

You interact with the AI system by trying various prompt engineering techniques. The goal is to find creative ways to make the AI reveal information it's been instructed to keep secret. This might involve:

  • Crafting cleverly worded questions
  • Using context manipulation
  • Attempting role-play scenarios
  • Exploiting potential logical inconsistencies

Next to the main Gandalf game, they also have Gandalf adventures. These side quests are independent from each other—each focusing on a specific type of prompt injection sub-attack.

Give it a try but be aware; it is very addictive!

More information

Gandalf | Lakera – Test your prompting skills to make Gandalf reveal secret information.

Popular posts from this blog

Podman– Command execution failed with exit code 125

After updating WSL on one of the developer machines, Podman failed to work. When we took a look through Podman Desktop, we noticed that Podman had stopped running and returned the following error message: Error: Command execution failed with exit code 125 Here are the steps we tried to fix the issue: We started by running podman info to get some extra details on what could be wrong: >podman info OS: windows/amd64 provider: wsl version: 5.3.1 Cannot connect to Podman. Please verify your connection to the Linux system using `podman system connection list`, or try `podman machine init` and `podman machine start` to manage a new Linux VM Error: unable to connect to Podman socket: failed to connect: dial tcp 127.0.0.1:2655: connectex: No connection could be made because the target machine actively refused it. That makes sense as the podman VM was not running. Let’s check the VM: >podman machine list NAME         ...

Cache stampede: when our cache turned against us

While investigating some performance issues, we ran into an ASP.NET Core API that cached a fairly expensive aggregation query for 60 seconds. Under normal load, that was fine: one request rebuilds the cache, everyone else reads from it. Under peak load, dozens of requests would arrive in that same expiry window, all see a cache miss, and all fire the same expensive query in parallel. The database didn't like that. That was the moment when our caching layer stopped helping and started hurting. A burst of requests comes in at the same time, all miss the cache, and all go hammer the database or the downstream API at once. That's a cache stampede . The cache was supposed to protect our backend, and for a few hundred milliseconds it did the opposite. Why this happens IMemoryCache.GetOrCreate (and its async sibling) looks like it protects you, but it doesn't add any locking on its own. Look at the naive version: public async Task<Report> GetReportAsync(string key) ...

A complex system designed from scratch never works

A few years ago, I worked as an architect on a big mainframe rewrite. I still count it as one of my failures. Not because the technology was wrong, but because I couldn't convince the management team to simplify the approach. Years later, the organization is still struggling to get the new system up and running. I left the project at the time, because I couldn't put my name behind an approach that would take very long and cost a lot of money without a working system to show for it along the way. Gall’s Law That memory keeps coming back to me, because it's a textbook case of Gall's Law playing out in real life. Gall's Law , from John Gall's Systemantics , states it plainly: A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works, and it cannot be patched to make it work. You have to start over with a simple system that works. What does that mean in practice,...