Blog
Practical notes on shipping code that AI assistants helped write: what tends to go wrong, how to catch it, and what we learn while building vetkit.
GuideA practical checklist for reviewing AI-generated code before you mergeTen things to check in code an assistant wrote, from invented dependencies and browser-only auth checks to swallowed errors and tests that assert nothing. · 9 min read
GuideSlopsquatting: when your AI assistant installs a package that does not existCoding assistants sometimes invent package names that attackers can register. What the research shows and how to check a dependency in a minute. · 8 min read
GuideYour LLM API key is probably in your frontend bundleWhy VITE_ and NEXT_PUBLIC_ variables end up in the browser, how to check a build for a leaked OpenAI or Anthropic key, and the server route that fixes it. · 7 min read