Blog
Thoughts, updates, and insights from the Superagent team.
Three years later: AI can (now) defend AI
In 2022, Simon Willison argued that 'adding more AI' was the wrong fix for prompt injection and related failures. He was mostly right at the time. What people tried then were brittle ideas that either overblocked or were easy to trick. This post explains what has changed since, what has not, and why builders can now use AI to meaningfully defend their agents in production.
Introducing Superagent — Defend Your AI Agents in Runtime
Today, we are proud to announce Superagent — the runtime defense platform that keeps your AI agents safe from prompt injections, malicious tool calls, and data leaks.
Alignment Faking: The New AI Security Threat
The development of sophisticated Large Language Models has introduced alignment faking as a critical challenge to AI safety. This strategic deception fundamentally complicates traditional safety measures, necessitating robust technical countermeasures.
Announcing Support for Cursor Agent and OpenCode
Every developer has preferences. Some love Claude's reasoning approach. Others prefer Cursor's interface and workflow. But you shouldn't have to compromise on security just because you prefer a certain agent. VibeKit's universal agent support provides a consistent security and observability layer that works across all your preferred agents.
Introducing VibeKit CLI
Every time you run an AI coding agent, you're giving it direct access to your environment. That moment of hesitation before you let the agent execute commands? We solved that. VibeKit is the safety layer that should have existed from day one.
Introducing Dagger Local Sandboxes
VibeKit now supports Dagger-powered local sandboxes for completely local AI code execution with container isolation and zero cloud dependencies, providing maximum privacy and performance for AI coding workflows.
Join our newsletter
We'll share announcements and content regarding AI safety.