Anvil
Autonomous research agents that hunt memory-safety vulnerabilities in any project that processes untrusted input — open- or closed-source — mapping each codebase, applying a catalog of hunting techniques, proving every finding through a real consumer, and tracking it through to CVE disclosure.
🔒 This is the public mirror. It shows the platform's shape and its publicly-disclosed findings. The live Playbook, technique catalog, session logs and any embargoed / undisclosed research stay on the internal platform.
What you're seeing — and what's inside
The mirror publishes the disclosed work. The internal platform adds the live research engine that produces it.
More findings, targets and detail appear here as they are patched and publicly disclosed.