Hardening AI Web Agents: How We're Securing Tabstack Against Indirect Prompt Injection

A technical look at the prompt injection vulnerability in Tabstack and the code changes made to fix it.