Amazon AWS unveiled three “frontier agents,” including Kiro, independent agents designed to automate coding, security, and DevOps tasks with minimal human supervision.
AWS introduced three new frontier agents: Kiro, an independent agent; the AWS Security Agent; and the AWS DevOps Agent. Kiro studies a team’s existing code, practices, and workflows. It also picks up tasks from the backlog and works through them singularly, coding, updating, or refactoring as required.
According to AWS, Kiro maintains a “persistent context across sessions.” This allows it to manage long, multi-step programming tasks, indeed across multiple repositories, without constantly asking for instructions.
Meanwhile, the AWS Security Agent reviews code, runs security checks, and penetration tests, and offers fixes for vulnerabilities, automating what was formerly a slow, human-operated process.
The DevOps Agent handles deployment, performance, and structure checks, helping teams catch and fix issues quickly while keeping services running smoothly.
This shift marks a major change in software development. Rather than simple AI sidekicks, these frontier agents act like real team members, independent, environment-aware, and able to handle complex systems over days.
With frontier agents, AWS is pushing beyond short-lived adjunct tasks toward completely independent “digital workers.” The move reflects a broader industry trend of making AI integral to software development, security, and operations, not just a helper, but a collaborator.
