Show HN: SwarmWright, structured multi-agent AI defined in markdowns (www.swarmwright.com)

SwarmWright has emerged as an innovative self-hosted workbench for designing and managing AI agent swarms, allowing users to orchestrate agents using markdown-based configurations. This tool bridges the gap between total automation and manual coding by providing a structured yet flexible environment where AI agents can autonomously pursue objectives while enabling users to maintain oversight. Each agent can perform actions, spawn sub-agents, and make decisions, all within a defined topology where connections and permissions are clearly stated, preventing unauthorized actions and ensuring accountability. The significance of SwarmWright lies in its design philosophy that combines expressiveness with safety, making it accessible for both technical and non-technical users. Each agent operates with a clear "constitution," providing transparency about its functions and decisions, which is particularly beneficial for compliance and debugging purposes. With the ability to run Python scripts, query databases, and seamlessly switch between AI model providers, SwarmWright empowers organizations to scale their AI workforce while keeping data local and secure. Its real-time dashboard allows users to monitor agent activities and immediately address issues, making it a powerful tool for those looking to leverage multitudes of AI agents effectively.