Show HN: Autonomy – Self-Harness/Self-Directed AI Agent Core Under Development (autonomy-landing-page.vercel.app)

A new AI agent core named Autonomy has been introduced, designed for self-directed task execution without human intervention. At its core is the AgentLoop, which autonomously carries out tasks through a looped process that evaluates and selects actions using a skill library and advanced machine learning techniques. Each execution cycle, or "Turn," involves choosing relevant skills, generating action paths, and assessing outcomes, with clear termination criteria to prevent endless loops. This system also features a risk management framework and maintains an Event Sourcing model for complete lifecycle documentation. The significance of Autonomy lies in its potential to enhance AI capabilities in multi-task environments, allowing for greater autonomy in complex problem-solving scenarios. With features such as dynamic skill selection, automated learning proposals, and a robust approval policy for actions, this framework could drastically improve efficiency and adaptability in AI systems. By automating the generation of procedural skills and enabling a rich user interface for interaction, Autonomy positions itself as a pivotal development for the future of self-directed AI agents.