Tools for AI auditors and
AI safety/security researchers.

Our mission is to give auditors, both internal and third-party, and AI safety/security researchers superior tools to understand AI agent activity, identify harmful or misaligned behavior, and assess whether control measures are good enough to prevent harm. Our tools are open by default, designed to run in the user's own environment, and work for time-constrained audits.

Talk to us Read our 2026 Theory of Change

Working with on Control Tower
used by , which maintains a public fork of our proxy

what we build

Our work addresses two problems: investigation and testing.

Investigation

The problem

It is difficult to reliably work out what agents did. The evidence is too large, split across coordinating swarms of agents, sometimes deliberately obscured, and possibly incomplete or altered.

What we build

Tools to analyze agent activity, verify log integrity, and identify harmful behavior.

Testing

The problem

Control evaluations don't reflect how agents are deployed inside the labs. They run outside a realistic deployment/harness environment, and sabotage examples are either too obvious or too harmless. As a result, controls appear stronger than they actually are.

What we build

Tools to red-team AI lab systems and controls in realistic deployment environments, allowing auditors to assess whether safety controls work.

We will develop these capabilities with direct feedback from auditors (e.g., Faculty) and AI control researchers (e.g., Redwood Research). Over time, we aim to improve agent incident detection and measure how well controls perform across labs and models.

Tools

Luthien Proxy and Control Tower

→

Our 2026 Theory of Change

Threat model, approach and risks

→