Agent-Ops 0.4.0: AI proposes, human decides, program executes

Sergey Zhitinsky, founder of Git in Sky, has introduced version 0.4.0 of Agent-Ops, an open methodology for collaboration between engineers and AI agents in IT infrastructure. Launched earlier this year, the project is now developed by a consortium of companies, transitioning from a single-company initiative to an industry-wide standard. The core concept of Agent-Ops focuses on restructuring workflows: AI handles problem investigation and solution proposals, while humans retain final approval authority and accountability. The 0.4.0 release is now available on GitHub and GitVerse as a public regulatory candidate. The author invites the professional community to contribute and discuss the framework, emphasizing that the methodology aims to adapt the IT industry to new automation realities, specifically addressing how responsibility is managed during infrastructure changes and incident resolution.