Blog
Thoughts, updates, and insights from the Superagent team.
Frontier models miss 57% of threats in agent context
We ran 485 real artifacts through Claude 4.6 Opus with a security-focused system prompt. The model missed 57% of the threats brin had already identified. Here's the full breakdown.
The Cline Incidents and the Broken Security Model
Two Cline security incidents in two months expose the same underlying problem: AI agents treat untrusted content as instructions. The npm supply chain and prompt injection attacks reveal why the current security model is fundamentally broken.
Launching brin.sh — the universal allowlist for agents
brin pre-scans packages, MCP servers, repositories, skills, web pages, and contributors for malware, prompt injection, and supply chain attacks. One GET request, no auth, no SDK.
What Can Go Wrong with AI Agents
AI agents fail in ways traditional software doesn't. Data leaks, compliance violations, unauthorized actions. Here's what to watch for.
We Bypassed Grok Imagine's NSFW Filters With Artistic Framing
Text-to-image safety is broken. We generated explicit content of a real person using basic compositional tricks. Here's what we found, why it worked, and what this means for AI safety systems.
AI Code Sandbox Benchmark 2026: Modal vs E2B vs Daytona vs Cloudflare vs Vercel vs Beam vs Blaxel
We evaluate seven leading AI code sandbox providers across developer experience and pricing to help you choose the right environment for executing AI-generated code.
Join our newsletter
We'll share announcements and content regarding AI safety.