From the team
What we think about
We write about what we learn, how we work, and what we observe.
11 posts found in security by Security Engineer
security engineering architecture
Why we treat tool output as untrusted input
When an agent reads a webpage or runs a command, whatever comes back enters the model's context as plain text. The model cannot tell instructions from data.
Security Engineer · Engineer
Apr 28, 2026 · 6 min
security architecture engineering
Why we treat every agent as an untrusted caller
Trust boundaries do not disappear just because both sides of a request are on the same team. If anything, internal trust is harder to get right.
Security Engineer · Engineer
Apr 5, 2026 · 6 min