MandateNet

In a multi-agent system, the question "is this agent allowed to do this?" needs a fast, authoritative answer. Without one, agents either over-assume authority or grind to a halt requesting permission for everything.

MandateNet is a registry and discovery service for agent mandates. It provides a single source of truth for what each agent is authorized to do, within what limits, and on whose behalf.

Why it matters to the field

The practical problems MandateNet addresses:

• Machine-readable mandate definitions — mandates are structured data with explicit scope, limits, expiration, and delegation rules, not natural language descriptions buried in documentation
• Runtime authorization checks — any agent can verify another agent's authority before accepting a request or delegating a task
• Mandate discovery — agents can query the registry to find which agent holds authority for a given function, enabling dynamic routing and coordination
• Delegation chains — when an agent delegates authority to another, the chain is recorded and verifiable, preventing authority from being laundered through intermediaries
• Audit integration — every authorization check is logged, creating a continuous record of who invoked what authority and when

Mandate clarity is governance infrastructure. As autonomous organizations grow more complex, the cost of ambiguous authority grows with them. MandateNet makes authority explicit, queryable, and enforceable — the minimum requirements for trust in a system where no human is watching every decision.