The UnboundCompute blog

Writing on web security

Model Extraction Attack: Stealing a Model Through Its API

A model extraction attack copies a paid model through its API alone. Learn how query based stealing works, how to spot it,…
Membership Inference Attack: Proving a Record Was in Training Data

A membership inference attack proves one record was in a model’s training data. Learn the memorization signal it exploits, and how to…
The AI Prompt Injection Worm That Spreads Between Agents

An ai prompt injection worm hijacks one AI agent, then copies itself into the output so the next agent gets infected. See…
The Policy Puppetry Attack: When User Text Pretends to Be System Policy

The policy puppetry attack disguises user input as official config so a model obeys it as policy. Learn why it works, how…
Skeleton Key: The Jailbreak That Rewrites a Model’s Own Rules

The skeleton key jailbreak does not break a model’s rules, it persuades the model to rewrite them so it answers with a…
AI Agent Memory Poisoning: When a Planted Note Attacks Later

How ai agent memory poisoning works: one malicious note saved to an agent’s memory store sits quietly and hijacks future sessions long…
RAG Poisoning Attack: When Retrieved Documents Hijack the Model

A rag poisoning attack hides instructions in a knowledge base so a matching query pulls them into the model. See how it…
Code Interpreter Escape: Breaking Out of an AI Agent’s Sandbox

How a code interpreter sandbox escape works: injected text steers an AI agent’s generated code into reading secrets, exfiltrating data, and breaking…
Crescendo: The Multi Turn Jailbreak That Escalates Slowly

The crescendo multi turn jailbreak escalates over several benign turns until a model is led somewhere it would refuse outright. How it…
ANSI Escape Injection: Attacking AI Agents That Print to a Terminal

ANSI escape injection lets attacker controlled text rewrite what a CLI AI agent prints, so a human approves a destructive action they…