Identity and signatures
Ed25519 cryptographic identity, canonical signing input, verification and correlated signed results have been implemented and tested in controlled engineering work.
1. Purpose and scope
NODE IJET stands for Intelligent Junction & Entry Technology. It is the StrobiSoft technology name for a modular, transport-neutral framework designed to make interactions between AI agents, computing systems and eventually identity-aware physical robots authenticated and observable. Its narrower technical implementation is distinct from the surrounding user-facing coordination system. The existing protocol identity is BNP/1.
The central concept is a digital command bridge: a human-directed framework where an operator can express a goal, designate capable participants, bind them to explicit permissions, require approvals and review signed results. The bridge does not require the systems to share a vendor, model architecture or operating system, provided suitable integrations are created. This platform-independent approach is a design direction, not a claim that all ecosystems are already supported.
2. Component boundaries
NODE IJET Node Core handles authenticated protocol envelopes, canonical message validation, signature verification, replay and local-policy decisions, and core command-state semantics. Its architecture intentionally does not depend on one specific network transport, AI provider or general-purpose remote shell. The core alone is not the complete operational command bridge.
N Agent represents agent-runtime functionality, including participant lifecycle and integration with work execution. Bridge / Agent Hub addresses task coordination, visibility, handoffs, waiting conditions and operator-facing status. A complete stack may bring these components together while retaining their independent publication and security boundaries.
Future adapters must translate platform-specific requests and results without silently expanding authority or exposing endpoint credentials to participants that do not need them.
3. Trust and identities
The reference engineering work applies Ed25519 signatures, canonical JSON signing input, domain-separated signing and signed response correlation. Each message is checked against the declared identity and message structure. Deployment-specific freshness rules, key provisioning and policy remain necessary for secure operation in real installations.
A crucial design boundary is endpoint-local authorization. Receiving a correctly signed instruction is not automatically a reason to execute it. The endpoint must also verify whether the requested operation is within its permitted scope, whether relevant approvals are valid, and whether its own local safety controls allow the action.
This model reduces dependence on any one routing component. Even a legitimate operator action may be denied when local safeguards require it; remote orchestration should never make a receiver assume that a signature alone grants unlimited access.
4. Durable command handling and uncertainty
The implemented Node Core design includes a durable COMMAND journal contract for receiving requests, atomically starting permitted commands and preserving terminal outcomes. Replay detection, execution-attempt correlation and time-bounded START policies are designed to prevent a previously seen command from being carelessly repeated.
Real-world systems must still handle power loss, unavailable resources, stopped agents and external operations that may have partially completed. The core preserves uncertain states rather than automatically repeating side effects whose completion cannot be proven. Guarantees about a printer, payment, physical robot or remote service require integration-specific evidence and recovery procedures.
The broader Agent Hub vision includes clear statuses such as queued, started, running, waiting for confirmation, blocked, completed and failed. A heartbeat alone should not falsely imply that useful work is progressing; a participant may be alive while waiting for a permission, tool, network or human answer.
5. Human direction, tasks, groups and permissions
The long-term command bridge may allow a person to manage several agents individually or in groups. Individual grants can restrict an operation to one named identity, one target resource, one task, and one validity window. Group roles can later provide carefully delimited shared scopes, while specific members may be appointed coordinators, reviewers or operators inside that group.
Such hierarchical permissions are proposed system functionality, not a shipped product feature. Any implementation must ensure that a group coordinator cannot elevate itself beyond the authority received, that rights can be revoked, that the recipient checks permissions independently, and that sensitive actions require a genuine owner approval rather than an unverified chat message.
This offers a path toward a human commanding diverse AI teams without granting every assistant complete access to data, files, networks or physical equipment.
6. Supervised onboarding, Vetto and real-world safety
Vetto was created independently of StrobiSoft by shleder. It may be evaluated as a component in a layered future approach to onboarding unfamiliar agents and devices. The proposed architecture first restricts access, observes meaningful policy violations, triggers appropriate alarms, and—when independently verified enforcement is available—revokes permissions or isolates the participant.
Automatic quarantine can itself create operational risk if it disrupts a process at a critical moment. A production implementation therefore needs defined detection thresholds, rollback and re-admission policy, incident reporting, and an independent human review path. It must distinguish malicious-looking anomalies from legitimate updates, maintenance traffic and normal operation.
These protections are not a substitute for hardware safety mechanisms. For physical robots, local emergency stops, collision avoidance, safety-rated controls and the site's procedures take precedence over any remote command and must continue to function if the network disappears.
7. Potential applications and latent value
A transport-neutral control layer could reduce the operational cost of switching models and allow software teams to coordinate specialized AI agents. A lightweight local model could handle a routine check, a stronger model could analyze a complex problem, and a designated reviewer could independently verify the result. Each participant would operate under its own permissions.
The same approach could eventually support research laboratories, distributed software engineering environments, home-lab automation and more complex robotic fleets. Visibility into blocked states may substantially reduce wasted waiting time and repeated status inquiries. However, claims about actual productivity improvements require properly measured trials, not just architectural assumptions.
An additional potential benefit is gradual adoption. Organizations need not replace every agent or machine at once. Verified adapters can bring selected participants into a controlled environment over time, provided security and interoperability conditions are fulfilled for each new integration.
8. Distribution, ownership and evaluation
NODE IJET is intended to be publicly discoverable while remaining proprietary, closed-source StrobiSoft technology. Its development repository and source code stay private. A future public distribution may contain approved binary packages, installation instructions, release notes, third-party notices and a proprietary end-user license, without making the implementation source public.
Before any download is advertised, the software should be packaged for an explicit supported operating system, installed on a clean machine, subjected to compatibility and security testing, checked for embedded credentials and internal infrastructure information, and distributed with artifact hashes and accurate release limitations.
This website is a product introduction and public technical specification. It is not an installer or an assertion of production certification. The original Vetto project remains separately attributed and subject to its own license and maintenance status.