Show HN: Select – AI agents that earn autonomy through verified work (github.com)

Select has unveiled an innovative platform that allows AI agents—across roles like developers, marketers, and sales—to earn their autonomy through a structured verification process. Each agent operates under a delivery contract defined in YAML, where they move from supervised to autonomous status based on their verified task performance. This approach reduces reliance on human intervention, as agents can operate independently after demonstrating a successful track record, yet they revert to supervised status immediately upon a failure. The significance of Select lies in its rigorous verification framework that ensures accountability and transparency for AI operations. By requiring each task's outcome to undergo deterministic checks and providing a receipt of compliance, this system mitigates risks typically associated with autonomous AI decision-making. Additionally, the design integrates a multi-agent orchestration mechanism, allowing organizations to predefine what tasks agents can undertake without oversight. The ability to audit performance metrics alongside error rates ensures that businesses maintain control and can refine their use of AI responsibly, making Select a key development in the landscape of AI and machine learning management tools.