This whitepaper introduces Lex Autonoma, a structural framework designed to enable autonomous AI agents to participate legitimately in digital economies. It addresses the institutional gaps preventing scalable agentic commerce: the lack of recognized identity, enforceable delegation, and shared interaction protocols for non-human actors. The core contribution is the Delegated Agency Framework (DAF), an implementation-ready architecture built on three irreducible pillars: Synthetic Identity (verifiable credentials for non-human actors), Delegated Agency (bounded autonomy with explicit scope constraints), and Contractual Physics (mechanical rules for fair, predictable multi-agent interaction). These pillars are operationalized through Machine-Legible Legal Primitives (MLLPs)—modular components that bridge legal intent and executable logic while ensuring verifiability, revocability, and jurisdictional compliance. Drawing on real-world examples such as the 2025 Amazon–Perplexity dispute, the whitepaper diagnoses current impasses—fragmentation, distrust, and exploit vulnerabilities—and argues that the DAF represents the minimal interoperable substrate any viable agent economy requires. It includes detailed taxonomies, protocol flows, security models, liability classes, and implementation pathways, generalizing beyond commerce to domains including finance, procurement, and IoT coordination. This framework is not a policy proposal but a minimal viable architecture, emphasizing predictability over restriction to make autonomy insurable and scalable. Appendices provide baseline MLLP specifications, system diagrams, and dependency graphs for developers and researchers.