Projects and adventures.
Set up Claude Code as a local assistant for your homelab: SSH setup, reusable prompts as slash commands, and automated security audits without manual back-and-forth.
An overview of the BankGPT challenge on TryHackMe: prompt injection techniques, LLM security, and context manipulation in a simulated banking chatbot.
An overview of prompt injection vulnerabilities in LLMs and how carefully crafted prompts can lead to the disclosure of protected information from model configuration.
