Veria Labs provides continuous AI-driven penetration testing that integrates into development workflows. The platform uses AI agents to systematically analyze codebases and applications for security vulnerabilities. The system operates in three main phases: it connects to git repositories where users define scope and security context, conducts deep analysis of codebases on setup and continues analysis on every pull request through CI/CD pipeline integration, and generates exploit proof-of-concepts that run against staging environments to validate real vulnerabilities. For each vulnerability discovered, the platform provides suggested patches that can be applied directly to the codebase. The AI agents analyze entire application scope including endpoints and logic to provide comprehensive coverage. The team behind Veria Labs consists of competitive hackers who have found vulnerabilities in major platforms and AI tools. The platform has discovered vulnerabilities in various systems including Cal.com, Claude, Gemini, and MCP. The service is backed by Y Combinator and focuses on eliminating the traditional tradeoff between development speed and security by providing continuous automated security testing rather than periodic manual penetration tests.