URDev4ever/LATT ? reverse-engineered prompt

Build me a small, clean toolkit repo for testing LLM apps for prompt injection, jailbreaks, context leakage, tool misuse, and workflow abuse.

I want it to feel like a practical security research guide, not an automated scanner. The main content should be a structured checklist of attack ideas, testing strategies, and mental models that help someone think through how an LLM system can break. Also include an interactive choose your own exploit style text guide that walks a tester through different branches and decisions while they explore a target.

Make the README easy to understand for security researchers and builders, with a short explanation of what the toolkit is, what it is not, and how to use it. Keep the writing clear and organized, and include an English version plus a Spanish version if that fits. If you need current references or best practices for LLM security wording, look them up online first.