<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> This prompt provides a comprehensive blueprint for constructing a maximally capable, self-improving agentic operating system. It targets developers and researchers aiming to build AI systems that can handle a vast array of computer-based tasks, from software engineering to complex project management. The core benefit is a system that learns from every interaction, continuously enhancing its own capabilities and autonomy over time.

How It Works

The system is designed around a closed-loop paradigm: goal -> task graph -> execution -> verification -> memory update -> visibility -> learning. It emphasizes a robust systems-engineering approach, prioritizing reliability, observability, and measurable progress. Key architectural choices include a strong single-agent baseline, explicit task graphs, durable file-based memory, and rigorous verification layers. The prompt details a capability acquisition ladder and a momentum engine to ensure continuous, compounding improvement.

Quick Start & Requirements

To begin, copy the provided prompt and paste it directly into a compatible AI coding agent (e.g., Claude Code, Codex, Cursor). The agent will immediately start building the system based on the prompt's instructions. The primary requirement is a capable AI agent that can interpret and execute complex, multi-step instructions. No specific installation or dependencies are listed for the prompt itself, beyond the underlying AI agent.

Highlighted Details

• Broad Task Handling: Capable of software engineering, debugging, research, automation, data analysis, and even company operations.
• Self-Improvement: Designed to learn from every task, iteratively enhancing its own prompts, skills, tools, workflows, evals, and architecture.
• Reliability Focus: Emphasizes deterministic workflows, validation gates, specialized harnesses, and compensating actions for critical business processes.
• File-System First: Projects are treated as durable operating systems, with files serving as the canonical memory and state, ensuring continuity across sessions.

Maintenance & Community

The prompt itself serves as the evolving specification. Details on specific contributors, community channels, or a formal roadmap beyond the prompt's structure are not provided in the README snippet.

Licensing & Compatibility

The prompt is designed to be runtime-agnostic, adaptable to various AI agents and environments. No specific open-source license is mentioned in the provided text.

Limitations & Caveats

This is a detailed prompt intended to guide an AI agent in building a system, not a pre-built application. Its effectiveness is entirely dependent on the underlying AI agent's ability to interpret and execute complex instructions, manage files, and perform software engineering tasks. The prompt's length and detail may pose a challenge for less capable agents.