Mechanism

The disclosed substrate carries the agent, not a packet, and that is what makes operation over disconnected and intermittently connected networks possible. Each agent is a cryptographically self-contained operand comprising a unique identifier, a payload, a memory field, a transport header, and a cryptographic signature. Because the agent carries its own execution context, a node can receive, validate, store, and later forward it without any persistent session, source-address routing, or transport-layer continuity. The substrate is described as deployable over delay-tolerant mesh architectures and across asynchronous or disconnected environments, including edge deployments and interplanetary communication links.

When a node receives an agent, it reconstructs the canonical serialization of the agent's unique identifier, payload, transport header, and memory field, and validates the cryptographic signature against the sender's public key. If validation fails, the validation layer discards the object and records the rejection outcome. If validation succeeds, the node parses the transport header and memory field to determine whether the agent is processed, forwarded, cached, or discarded. This same evaluation applies regardless of how long the agent has been in transit, so an agent may be validated and processed even after long propagation delays.

Why Carriage Works Without a Session

The carry-and-forward behavior is a consequence of putting the governing state inside the agent rather than in the network. The transport header defines propagation constraints such as time-to-live, trust radius, and semantic class, and these values determine admissibility at the current node and influence whether the agent is processed, forwarded, cached, or discarded. The memory field contains verifiable lineage, access logs, and policy references, and these records carry the agent's behavioral history and governance context into every node it contacts. No external session management, centralized controllers, or pre-configured address registries are required.

Because the agent provides the authoritative and sufficient basis for secure execution, a node that holds an agent during a period of disconnection does not need to coordinate with any other party to decide what to do with it. The transport header's constraints and the memory field's policy references are evaluated locally. The disclosure states that embedded policy enables secure operation in disconnected or intermittently connected networks such as IoT or interplanetary systems.

Caching, Rerouting, and Delay-Tolerant Propagation

The substrate is transport-agnostic. Agents may be serialized as structured payloads, transmitted without alteration, and reconstructed at the receiving node, and their internal structure and behavioral determinism are preserved regardless of connection lifetime, packet ordering, or relay topology. Within this model, nodes may cache unresolved agents, forward them via delay-tolerant paths, or distribute them over broadcast overlays as conditions dictate. Regardless of transport behavior, each agent remains a fully portable, self-contained behavioral unit.

Each layer of the protocol stack that processes the agent appends a corresponding execution trace to the memory field. These traces include both local decisions, such as routing outcomes or rejection causes, and network-wide feedback such as system health, cache status, or propagation entropy. The selected next hop is appended to the agent's memory trace, and the agent is forwarded accordingly. Downstream nodes reference this accreted trace to validate, replay, or audit prior execution outcomes, ensuring trust continuity and semantic audit across asynchronous or intermittently connected environments.

Trust-Scoped Forwarding at Each Hop

Forwarding decisions during carriage use the same dynamic routing protocol that governs live transmission. The dynamic routing protocol is a memory-aware, behavior-sensitive routing layer that directs transmission based on trust scope, access history, policy constraints, and dynamic system health rather than static addresses or hop-count heuristics. Upon receiving an agent, the node examines the access log to identify recent execution history associated with neighboring nodes, including success rates, policy violations, and responsiveness, and assigns dynamic trust scores to routing candidates. Nodes failing trust or policy thresholds are excluded.

Agents may be classified as forwardable, suppressible, or urgent, and these classifications may originate from the initial sender or be updated by intermediate nodes in response to propagation failures or system signals. Agents violating time-to-live, trust scope, or other constraints are dropped or quarantined, and these decisions and their justifications may be appended to the agent's memory trace. A node may also append network-health-derived observations to the memory field of an active agent, enabling downstream nodes to inherit awareness of recent network conditions as part of their routing context.

Stateless Edge Carriers

The substrate supports fallback execution in fully stateless environments. Nodes configured without persistent memory rely entirely on the agent's embedded data for trust evaluation, quorum participation, and policy enforcement. This capability allows devices with limited resources or transient uptime, including IoT devices, ephemeral containers, or anonymized relays, to participate in substrate behavior without requiring full-stack deployment or long-term data retention. In such contexts, the agent itself provides the authoritative and sufficient basis for secure execution.

In edge deployments such as mobile devices, remote sensors, or IoT nodes, minimal substrate configurations allow these devices to receive and forward agents, evaluate routing and trust constraints, and optionally cache or discard data based on local policy. These environments typically operate a dynamic routing protocol and a simplified semantic memory layer, often configured in stateless mode to preserve limited resources. Even so, such nodes remain fully interoperable with more capable peers while still enforcing trust-scoped routing and policy-aligned behavior.

Integrity Across Carriage

The integrity of a carried agent does not depend on trusting the intermediate nodes that held or relayed it. Each agent includes a cryptographic signature generated over a canonical serialization of its unique identifier, payload, transport header, and memory field, and validation of the signature at each hop preserves authenticity, continuity, and policy alignment before the agent is executed or forwarded. This mechanism prevents unauthorized mutation, tampering, or policy subversion.

Within the memory field, each trace entry is independently signed by the node that generated it and chained using cryptographic hashes, ensuring chronological ordering, auditability, and non-repudiation. A consensus node can later evaluate whether an agent's current state is derivable from a trusted origin by inspecting its mutation lineage, which provides a substrate-level defense against unauthorized forking, out-of-scope overrides, or stale mutation replays. The result is that a node which receives an agent after an arbitrary period of carriage evaluates it under the same signature verification, lineage, and policy checks it would apply to any other agent.

Deployment Contexts

Because agents carry all necessary execution context, including policy references, mutation proposals, quorum metadata, and routing constraints, the system functions in asynchronous and delay-tolerant environments. The disclosure identifies edge deployments such as mobile devices, remote sensors, and IoT nodes operating minimal stateless configurations, high-availability core nodes running the full protocol stack, and federated or cross-domain deployments that operate across administrative boundaries without shared infrastructure or synchronized ledgers. The architecture supports evolutionary deployment, in which a node may begin as a stateless router and progressively adopt additional protocol layers as its role or resources expand, without reconfiguring identity or coordination logic.

The substrate may be implemented atop legacy transport layers including TCP/IP, HTTP, WebRTC, or delay-tolerant mesh architectures, and may be deployed incrementally across environments ranging from edge devices to interplanetary networks. This compatibility model enables integration with existing infrastructure without protocol replacement, and supports hybrid deployments in which substrate-native nodes interoperate with legacy clients.

Disclosure Scope

United States Patent Application 19/366,760 discloses the memory-native protocol substrate in which each agent is a self-contained operand carrying a unique identifier, a payload, a memory field with verifiable lineage, access logs, and policy references, a transport header, and a cryptographic signature, and in which nodes process, cache, forward, or discard agents based solely on agent-resident content. The disclosure encompasses operation over delay-tolerant and intermittently connected networks, the caching of unresolved agents and their rerouting via delay-tolerant paths or broadcast overlays, trust-scoped forwarding under the dynamic routing protocol with time-to-live and trust-radius constraints in the transport header, per-hop signature verification and hash-chained trace continuity, and stateless edge deployment on mobile, sensor, and IoT nodes.

The disclosed scope contemplates deployment across edge, core, and federated configurations, over transports including TCP/IP, HTTP, WebRTC, mesh relay, and delay-tolerant networking, and in asynchronous or disconnected environments up to and including interplanetary communication links. Implementations falling within this scope, regardless of transport choice, device class, or deployment topology, are within the disclosure.