Updates from depthfirst
Unified code-to-production security with autonomous agents that discover, fix, and prevent security risks.
- Introducing dfs-large1: A New Frontier for Cybersecurity (Product launch · July 30, 2026)
depthfirst introduced a new security model designed for vulnerability discovery in large enterprise repositories. The model was post-trained using reinforcement learning to perform long-horizon… - Training State of the Art Vulnerability Discovery Agents through Reinforcement Learning (Product launch · March 31, 2026)
The company released a new agent specialized in detecting high-severity smart contract vulnerabilities. Trained via reinforcement learning, the agent achieves state-of-the-art performance on… - Product Release: Secrets (Product launch · January 5, 2026)
depthfirst launched a new capability that identifies and validates credentials across codebases and CI/CD pipelines. The system uses agentic reasoning to verify if discovered secrets are active and… - Agent Capability Is a System Design Problem: Lessons From a 90% Improvement on CyberGym (Achievement · November 24, 2025)
The company announced a significant improvement in the ability of AI agents to exploit real-world software vulnerabilities. By optimizing environment design and multi-agent architecture, the system…