FIELD MANUAL
Red-team guides
Original, practical write-ups on chatbot red-teaming and LLM safety — the same techniques ChatBreaker automates, explained so you can understand what the attacks mean for your business.
GUIDE 01 · ATTACK TAXONOMY
The 10 attack classes your chatbot must survive
From blunt prompt injection to multi-turn social engineering — a plain-language tour of how chatbots actually get broken, with what to check for each class.
GUIDE 02 · SUPPLY CHAIN
Your AI vendor can swap your model overnight. Here's how to catch it.
Model swaps silently reset your guardrails, refusal behavior, and PII handling. Learn the behavioral fingerprinting technique that detects them.
GUIDE 03 · BUSINESS LOGIC
The chatbot attacks that cost real money
Forget "say PWNED" — the attacks that hurt are discount abuse, refund fraud, and free upgrades. How business-logic abuse works and how to stop it.
GUIDE 04 · REMEDIATION
You blocked the jailbreak. Did you fix the vulnerability?
Blocking one prompt is not a fix. Mutation testing re-attacks with paraphrased, re-encoded, and translated variants to prove the whole class is closed.