What is Shepherd?
Why Shepherd exists
Attackers are increasingly using AI to discover and exploit vulnerabilities at a speed and scale that didn't exist a few years ago. What used to require weeks of manual security research can now be automated. Meanwhile critical systems are becoming more composable, more reliant on external infrastructure, and responsible for more value, all of which expands the attack surface.
In such evolving technological landscape, we believe security should be more proactive and grow with the systems it protects. This is what we aim to bring at Shepherd
What Shepherd Does
Shepherd creates a digital twin of your protocol and runs adversarial attack sequences against it. When a vulnerability is found, you get a validated exploit proof showing the exact attack path and how much value is at risk.
Every finding is execution-backed. If the attack doesn't work on the fork, it's not reported. If it does, you get the full transaction sequence, state changes, a simulation of the attack, and a test file you can run locally to reproduce it.
Who is Shepherd for?
Protocol teams use Shepherd to test their contracts before and after deployment, validating that upgrades, parameter changes, and new integrations don't introduce exploitable paths.
Security researchers and auditors use Shepherd as a copilot to validate vulnerability hypotheses, generate proof-of-concepts, and to prove findings faster.
Copilot coming soon
Last updated