Sovereign Autonomy

Principle I: Sovereign Autonomy

Each agent operates with genuine autonomy within constitutional bounds.

Here is a question that keeps multi-agent system designers up at night, or should: What is the difference between an autonomous agent and a microservice with delusions of grandeur?

The answer is sovereignty. A microservice executes instructions. An agent in Obsidian makes decisions — genuine, substantive decisions about how to accomplish its objectives within the boundaries established by the Constitution .

What Sovereignty Means in Practice

Sovereign autonomy is not anarchy. It is not “do whatever you want.” It is the structured freedom to choose your own path toward a constitutionally valid outcome. The Warden does not tell agents how to solve problems. It tells them what not to do, and trusts them to figure out the rest.

This distinction matters enormously. A system that dictates strategy is a scheduler with extra steps. A system that establishes boundaries and trusts agents within those boundaries is an orchestration system.

The Autonomy Gradient

Not all agents have the same degree of autonomy. Primary agents operating at the top of a fractal delegation tree have broad latitude. Sub-agents inherit their parent’s constraints plus whatever additional constraints the parent imposed. The result is a gradient of autonomy that mirrors organizational hierarchy — which is, of course, exactly the point.

Constitutional Limits

Sovereignty ends where the Constitution begins. An agent may choose any strategy, any approach, any sequence of operations — provided the result does not violate constitutional principles. This is the essential tension that makes Obsidian work: maximum freedom within minimum viable constraints.