Kiro has moved to general availability, introducing a series of features aimed at improving the reliability, governance and team-wide use of AI agents in software projects. The update follows several months of preview adoption, during which spec-driven development gained traction among engineering teams seeking more structured ways to integrate AI into coding workflows.
One of the most significant additions is property-based testing, which evaluates whether code generated with AI aligns with the behavioural requirements outlined in a project’s specifications. Unlike traditional unit tests that check a limited number of scenarios, property-based testing generates hundreds or thousands of test cases to uncover inconsistencies, edge cases or deviations from expected behaviour.
Kiro has also added checkpointing, allowing developers to revert to earlier stages of an agent’s actions without losing overall progress. The platform now supports multi-root workspaces as well, enabling development across multiple project folders — a capability relevant to teams working with multi-module or multi-repository structures.
In parallel, the company has introduced the Kiro CLI, bringing the platform’s agent capabilities — including workflow automation, error analysis and custom agent support — to the command line. The CLI uses the same steering files and configuration settings as the Kiro IDE, ensuring consistency between environments for distributed teams.
For organisational use, Kiro now supports sign-in through AWS IAM Identity Center, giving administrators centralised control over access, billing, usage limits and configuration settings. Support for additional identity providers is planned.
Kiro has also announced a global offer for startups, providing one year of access to Kiro Pro+ for eligible early-stage companies until 31 December 2025. The scheme can be combined with AWS Activate credits.
The latest release reflects a broader industry shift towards bringing structure, verification and oversight to AI-assisted development, as organisations work to balance productivity gains with reliability and governance needs.