For researchers, operators, investors and policy teams building in agentic commerce.
57 definitions catalogued. 17 protocols indexed. 4 published papers. One independent map, updated every Monday.
The intelligence exists. The communication exists. The economic infrastructure does not.
Agents can negotiate, but they can't pay. No protocol handles escrow, disputes, or refunds between two AI agents without a human in the loop.
An agent's track record dies with its session. There's no portable, verifiable history that follows an agent across platforms and transactions.
When an autonomous agent makes a bad trade, who's liable? When it colludes with another agent, who detects it? The rules don't exist yet.
10 components. Score 0-100. Portable via JWT. Sybil-resistant by design. The first quantitative reputation system built specifically for autonomous agent commerce.
Read on Zenodo →57 definitions. 6 categories. 2 fault lines. Why 'agentic economy' means at least two completely different things — and why the infrastructure for each is incompatible.
Read on Zenodo →Gödel, Tarski, and Rice walk into an LLM evaluator. Formal proof that semantic truth verification is computationally intractable — and an architecture that works anyway.
Read on Zenodo →57 definitions, zero convergence. Three historical analogs, four falsifiable tests. Why the agentic economy is a pre-infrastructural phase whose vocabulary will be selected by what gets built.
Read on Zenodo →The most comprehensive comparison of agent commerce protocols and payment standards. Who built what, what each assumes, and what's still missing.
See the full comparison →