ctf-ai-ml

Solve AI and machine-learning CTF challenges.

CTF security

What it does

Applies adversarial AI and model-analysis techniques to solve AI/ML capture-the-flag challenges.

When to use it

Use when solving CTF challenges involving model attacks, adversarial examples, prompt injection, inference, poisoning, or LLM jailbreaking.

It changes how the agent approaches AI/ML CTF challenges by supplying attack techniques and detailed supporting references.

This skill

challenge LLM endpoint

Write

Sends attack prompts to endpoints

torch, transformers, numpy, scipy, Pillow, safetensors, scikit-learn

Execute

Installs AI/ML Python packages

Requires Python 3.

Requires the torch Python package.

Requires the transformers Python package.

Requires the numpy Python package.