Sovereign-OS (V16 Pristine)

Sovereign-OS is a hyper-minimalist meta-framework for governing autonomous AI coding agents across projects. Built strictly on the Ponytail doctrine, it provides exactly what is needed to manage agents and absolutely nothing more.

Architecture

At its core, Sovereign relies on a single PowerShell orchestrator (sovereign.ps1) that enforces single-instance execution via an OS-level Mutex and manages configuration via sovereign.config.json.

The system's capabilities are derived entirely from Git Submodules and the Dynamic Asset Registry.

graph TD
    A[Sovereign Core] --> B(sovereign.ps1)
    A --> C(sovereign.config.json)
    
    B --> D[Modules]
    D --> E(no-mistakes)
    D --> F(codebase-memory-mcp)
    
    B --> G[Skills]
    G --> H(ponytail)
    G --> I(agent-reach)
    
    A --> J[Dynamic Asset Registry]
    J --> K(CI/CD)
    J --> L(Security)
    J --> M(Linting)

The Dynamic Asset Registry

Sovereign-OS does not vendor large external dependencies (like CI/CD pipelines, security scanners, or frontend frameworks) by default. Instead, it maintains a curated ASSET_REGISTRY.md.

Agents operating within Sovereign-OS read this registry and dynamically integrate tools using agent-reach only when a user task actively requires them. This ensures the core repository remains pristine and unbloated.

Governance

All operations within Sovereign-OS are bound by the Standing Directive located at .agents/AGENTS.md and the philosophical constraints of the ponytail skill. Proof of system integrity is strictly documented in AUDIT_LEDGER.md.