RunSybil is an AI-powered continuous attack surface assessment and validation platform designed to help security teams discover, test, and remediate vulnerabilities across their application attack surface. The platform operates across four core phases: - Map: Automatically discovers hidden endpoints and forgotten assets across cloud environments and third-party services to build a unified inventory of the attack surface, covering both internal and external assets. - Attack: Conducts automated offensive testing against the discovered attack surface with context-aware analysis. Features include transparent reasoning behind findings, verified results, complex attack path discovery, and variant hunting to identify related vulnerability classes. - Fix: Supports remediation workflows through vulnerability prioritization, remediation guidance, and mitigation recommendations to help teams address findings efficiently. - Test: Validates that fixes are effective through unlimited retests and the ability to rerun full attack campaigns. Provides reporting to demonstrate security posture improvements over time. The platform surfaces findings with severity indicators and tracks accepted risks, as shown in its vulnerability dashboard. It is built by a team with backgrounds from OpenAI, Meta, Mandiant, NCC Group, and Trail of Bits.