Cognitive Mesh is an enterprise-grade AI transformation framework designed to orchestrate sophisticated, multi-agent cognitive systems. Our mission is to provide a secure, compliant, and philosophically-grounded platform that enables organizations to build, deploy, and govern advanced AI capabilities with confidence and transparency.
The platform integrates spectrum-adaptive intelligence, a complete NIST AI RMF compliance system, and a foundational Zero-Trust security architecture, making it the most advanced and reliable system for regulated and mission-critical enterprise environments.
The Cognitive Mesh architecture is built on four unwavering pillars that ensure robustness, security, and scalability.
- Layered Hexagonal Architecture: A clean, decoupled design separating core domain logic from infrastructure. Each of the 5 layers (Foundation, Reasoning, Metacognitive, Agency, Business Applications) has a distinct responsibility, communicating through well-defined ports and adapters.
- Zero-Trust Security by Default: Security is not an afterthought; it is the foundation. Every request, whether internal or external, is authenticated, authorized, and encrypted, enforcing the principle of least privilege across the entire mesh.
- Ethical & Legal Compliance: The framework is governed by a comprehensive ethical and legal compliance system, ensuring all operations align with global standards like the GDPR and the EU AI Act, and are grounded in established philosophical principles.
- PRD-Driven Development: Every component, feature, and architectural decision is guided by a comprehensive portfolio of Product Requirement Documents (PRDs). This ensures systematic, transparent, and priority-driven development.
The Cognitive Mesh is organized into five distinct layers, each with a specific role in the cognitive processing pipeline. This separation of concerns ensures modularity, testability, and maintainability.
FoundationLayer: Provides the core infrastructure, including security, data persistence, audit logging, and communication protocols. It is the bedrock upon which all other layers are built.ReasoningLayer: Contains the cognitive engines responsible for various forms of reasoning—analytical, creative, ethical, and threat intelligence.MetacognitiveLayer: The "mind of the mesh," responsible for self-monitoring, continuous learning, performance optimization, and incident response.AgencyLayer: Home to autonomous agents that execute tasks, interact with tools, and carry out automated workflows based on the decisions and plans formulated by the other layers.BusinessApplications: The outermost layer, which exposes the mesh's capabilities to the outside world through controllers, APIs, and business-specific logic.
- .NET 9.0 SDK or later
- Visual Studio 2022 or VS Code with the C# Dev Kit
- (Optional) Azure Subscription for cloud-native feature development
-
Clone the repository:
git clone https://github.com/phoenixvc/cognitive-mesh.git cd cognitive-mesh -
Build the solution: This will restore all NuGet dependencies and compile every project.
dotnet build CognitiveMesh.sln
-
Run all tests: Verify that the entire system is functioning correctly.
dotnet test CognitiveMesh.sln
For advanced build, testing, and utility operations, please see the scripts in the ./scripts/ directory.
All development in the Cognitive Mesh is meticulously planned and tracked through a comprehensive portfolio of Product Requirement Documents (PRDs). This ensures every feature is well-defined, architecturally sound, and aligned with the project's strategic goals.
Our complete PRD portfolio, including implementation priorities, status, and dependencies, is tracked in the master document:
➡️ PRD Priority & Implementation Status
Contributions are what make the open-source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.
Please refer to our Contributing Guidelines for details on our code of conduct and the process for submitting pull requests to us.
This project is licensed under the MIT License. See the LICENSE file for more information.