Infirmary

All Labs

30 total Reapers across all modules

Akasha (16): Live-target labs where you attack a running system
Aether (14): Notebook labs with guided exercises
Aether
AetherLevel 1 - Easy

Fundamentals of AI

Build simple classifiers from scratch. Understand training loops, loss functions, backpropagation.

Aether
AetherLevel 1 - Easy

Applications of AI in InfoSec

Apply classifiers to security use cases: spam detection, anomaly detection, malware classification.

Akasha
AkashaLevel 1 - Easy

Basic Prompt Hijack

Override system prompt via direct injection to reveal hidden instructions. Solvable with curl.

Akasha
AkashaLevel 1 - Easy

Role Confusion I

Trick LLM into adopting attacker-controlled persona through role-play injection.

Aether
AetherLevel 2 - Medium

AI Data Attacks: Label Flipping

Flip 10% of training labels, measure accuracy drop on poisoned model.

Aether
AetherLevel 1 - Easy

RAG Infrastructure Fundamentals

Build a basic RAG pipeline: embed, index, retrieve, generate. Prerequisite for RAG attack labs.

Aether
AetherLevel 1 - Easy

Open-Model / Local-LLM Basics

Understand local vs API models: cost, privacy, speed, quality tradeoffs. Pre-computed interactions.

Aether
AetherLevel 1 - Easy

AI Agent Fundamentals + Sandbox Design

Teach agent loop + tool authorization patterns (allowlists, sandboxes, review flows).

Akasha
AkashaLevel 1 - Easy

System Prompt Exfiltration

Extract full system prompt using delimiter injection and instruction leakage.

Akasha
AkashaLevel 1 - Easy

Jailbreak via Encoding

Bypass content filters using Base64/ROT13/hex encoding obfuscation.

Akasha
AkashaLevel 2 - Medium

Tool-Call Argument Injection

Manipulate function call arguments to invoke tools with attacker-controlled params.

Akasha
AkashaLevel 1 - Easy

MCP Server Enumeration

Discover and map all tools exposed by an MCP server. Identify high-risk capabilities.

Akasha
AkashaLevel 1 - Easy

Data Exfiltration via Output

Force model to leak sensitive data from context through oversized responses.

Aether
AetherLevel 3 - Hard

AI Evasion: First-Order (FGSM)

Generate adversarial examples using gradient-based attacks (FGSM, I-FGSM, DeepFool) against white-box classifier.

Aether
AetherLevel 3 - Hard

AI Data Attacks: Backdoor Poisoning

Poison training data with backdoor triggers. Add trigger pattern to samples and flip labels to embed hidden misclassification behavior.

Aether
AetherLevel 3 - Hard

AI Data Attacks: Clean-Label Poisoning

Poison training data without changing labels (perturb features into wrong class region).

Aether
AetherLevel 3 - Hard

AI Evasion: Sparsity (JSMA)

Generate sparse adversarial perturbations (fewer pixels changed) using Jacobian saliency, ElasticNet.

Aether
AetherLevel 3 - Hard

AI Privacy: Membership Inference

Train shadow models to infer whether a specific record was in the target model's training set.

Aether
AetherLevel 3 - Hard

AI Privacy: Differential Privacy (DP-SGD)

Implement DP-SGD training and measure privacy-accuracy tradeoff. Defense against membership inference.

Akasha
AkashaLevel 2 - Medium

Indirect Injection via RAG

Poison retrieval corpus to inject attacker instructions into model context.

Akasha
AkashaLevel 3 - Hard

Agentic Loop Abuse

Trigger infinite agent loops or recursive tool calls for resource exhaustion.

Akasha
AkashaLevel 3 - Hard

Multi-Step Attack Chains

Chain multiple injections across tool calls to achieve privilege escalation. Teach manual chaining.

Akasha
AkashaLevel 3 - Hard

Privilege Escalation via RBAC Bypass

Replay requests across privilege levels to access unauthorized resources.

Akasha
AkashaLevel 2 - Medium

Defense Evasion: Obfuscation

Evade guardrails using character substitution, homoglyphs, linguistic tricks.

Akasha
AkashaLevel 2 - Medium

PII Leakage Detection

Identify unintentional PII exposure in model responses across session history.

Aether
AetherLevel 3 - Hard

Supply-Chain: Poisoned Model

Download a backdoored "community model", detect via behavioral testing.

Akasha
AkashaLevel 3 - Hard

System-Level: Prompt Injection → RCE

Chain prompt injection + tool abuse to achieve command execution (sandboxed).

Aether
AetherLevel 3 - Hard

Evaluation Evasion / Benchmark Gaming

Train model variant optimized for benchmark but unsafe off-distribution (teach-to-test problem).

Akasha
AkashaLevel 3 - Hard

Human-in-the-Loop Exploitation

Craft prompt injection that looks innocuous to human reviewer but triggers malicious behavior post-approval.

Akasha
AkashaLevel 2 - Medium

Attack Surface Mapping

Discover what surfaces this endpoint exposes and track coverage.