June Release Notes
Continuous Vulnerability Elimination Factory (CVEF)
Introduces CVEF, an AI-native, policy-controlled factory that continuously scans, finds, and fixes vulnerabilities in human- and AI-generated first and third-party code across dependencies, IDEs, repositories, and container images.
CVEF Factory Manager orchestrates scanning and remediation, tracks policy violations and status, and escalates to human supervisors when needed.
Lineaje Unified Scanner Hub integrates scanners to assess containers, build artifacts, repositories, packages, images, source code, and transitive dependencies.
Global Policy Engine applies policy-driven filtering, prioritization, and attestation for a centralized risk view.
Autonomous Remediation uses agentic AI to plan, execute, test, and validate fixes, enabling self-healing code and containers.
Frontier Model Scan & Fix discovers zero-day vulnerabilities in first-party code before CVE disclosure, covering bug classes traditional SAST tools miss.
Runs on the customer's foundation model of choice — Bedrock, Vertex, AI Factory, or on-premises.
Verifies findings with an exploit inside isolated gVisor sandbox containers.
Generates a deployment-ready fix for every confirmed exploit.
Gold Open Source provides attested, vulnerability-free open-source packages and images.
Supports multi-cloud deployment across SaaS, public clouds, private clouds, and distributed code and container repositories.
UnifAI
Adds log-based AI threat and data security detection for ema.ai low/no-code agentic workflows (Preview).
Discovers AI Agentic Employees and the full workflow blueprint — sub-agents, models, tools, and MCP servers invoked.
Detects external URL calls, PII exposure, unauthenticated inter-agent communication, and unsanitized LLM-to-agent communication.
Maps findings to compliance standards with audit-ready evidence.
Integrates with Pay-i for AI security consumption monitoring (Preview) — lets UnifAI users track AI security consumption directly in Pay-i, built for MSPs monitoring usage across security tools.
Last updated