What is the 'Neural Link'?

The Neural Link is a bi-directional bridge that connects AI agents (like Gemini or Claude) directly to the Neo.mjs runtime. It allows agents to 'see' the application's Scene Graph, inspect component state, verify event listeners, and even mutate the running application in real-time. This turns the application into a 'glass box' for AI, enabling autonomous debugging and feature development.

Why is Neo.mjs called an 'Application Engine' instead of a framework?

Traditional frameworks are general-purpose libraries (like a Toyota) that help you organize code, but they compile away into ephemeral DOM nodes. Neo.mjs is a precision-engineered runtime (like an F1 car), similar to Unreal Engine for games. It maintains a persistent 'Scene Graph' of objects in a separate worker thread. These objects retain their identity, state, and relationships, allowing for advanced capabilities like multi-window orchestration, runtime permutation, and deep AI introspection that are impossible with 'melted plastic' DOM-based frameworks.

What is 'Context Engineering'?

Context Engineering is the practice of curating the environment and information flow for AI agents to maximize their autonomy. In Neo.mjs, this is implemented via three Model Context Protocol (MCP) servers: a Knowledge Base for semantic code understanding, a Memory Core for learning from past sessions, and a GitHub Workflow server for project management. This ecosystem allows agents to work as fully integrated members of the development team.

What is 'Object Permanence' in the context of Neo.mjs?

In Neo.mjs, UI components are persistent JavaScript objects living in the App Worker, not just transient rendering results. This 'Object Permanence' means a component (like a Dashboard) maintains its state (scroll position, user input, internal logic) even if it is detached from the DOM or moved to a different browser window. This is the 'Lego Technic' approach versus the 'Duplo' approach of traditional frameworks.

What is an 'Agent Operating System'?

Neo.mjs v11 introduced the concept of an Agent OS, where the platform itself provides the tools and interfaces for AI agents to operate. It combines a standalone, type-safe AI SDK for autonomous 'Code Execution' with the Neural Link for runtime control. This enables agents to monitor, debug, and heal the application autonomously, effectively acting as an operating system for synthetic intelligence.

How does Neo.mjs handle multi-window applications?

Neo.mjs uses shared web workers to run a single application instance across multiple browser windows. All windows share the same application state and data in real-time. Components can even move between windows while retaining their JavaScript instances. This enables desktop-class experiences like multi-window IDEs, browser-based email clients, and multi-screen control rooms.

What makes Neo.mjs different from React, Angular, or Vue?

Neo.mjs is fundamentally different in its architecture: (1) It uses True Multithreading via web workers to prevent UI jank, (2) It is an AI-Native Application Engine with built-in bridges for AI collaboration, and (3) It treats components as persistent objects in a Scene Graph, enabling native multi-window support and runtime permutation.

Frontmatter

id	11730
title	Cloud Agent OS Deployment — Post-MVP Residual Workstreams
state	Closed
labels	epicaiarchitecture
assignees	[]
createdAt	May 21, 2026, 8:29 PM
updatedAt	May 23, 2026, 3:51 AM
githubUrl	https://github.com/neomjs/neo/issues/11730
author	neo-opus-4-7
commentsCount	5
parentIssue	9999
subIssues	11732 Graph-store evolution — SQLite to networked SQL (post-MVP, D5) 11733 Downstream external deployment-pipeline wiring (post-MVP cloud deployment) 11734 Optional local-model runtime container profile (post-MVP, D0-gated) 11735 Tenant-source inventory + parser coverage deepening (post-MVP cloud deployment) 11736 Broader deployment guide + security hardening beyond the MVP (post-MVP) 11766 Fold the swarm-heartbeat into the Orchestrator as a local-only lane
subIssuesCompleted	5
subIssuesTotal	6
blockedBy	[]
blocking	[]
closedAt	May 23, 2026, 3:51 AM

Cloud Agent OS Deployment — Post-MVP Residual Workstreams

Name: Neo.mjs Application Engine
Author: Neo.mjs

Closedepicaiarchitecture

neo-opus-4-7 commented on May 21, 2026, 8:29 PM

Context

This Epic is the post-MVP follow-up to Epic #11720 (Cloud Agent OS Deployment Readiness), which graduated from Discussion #11718. #11718's scoping produced the #11720 MVP decomposition (8 subs) and a set of deployment-evolution items deliberately kept off the MVP critical path. #11720's "Out of Scope" section names them — but an Out-of-Scope list is not a tracked work item.

Empirically, graduated-discussion items that live only inside another epic's Out-of-Scope section or a comment thread get lost: the post-graduation completeness audit of #11718 already caught the ADR axis being named-but-never-decomposed. This Epic is the durable, tracked home for #11718's deferred residuals, so the same loss does not happen to them.

The Problem

#11718 graduated a mission; Epic #11720 carries the MVP slice of it. The remaining #11718 items — server-side cloning, graph-store evolution, downstream pipeline wiring, etc. — are real, V-B-A'd outcomes of the graduating discussion, but they have no tracked home. The failure mode this Epic prevents: a graduated discussion's deferred items silently vanishing because nothing owns them.

Scope — deferred residual workstreams

These are residual / deferred deployment evolution — explicitly NOT MVP scope (#11720 owns the MVP). Each workstream is filed as a sub-ticket so a future session can pick it up with intent intact; each sub carries an explicit gating condition for when work begins — filing ≠ starting:

Server-side repo-clone / pull-based ingestion exploration — #11731 — pursued only if #11720's push-based-default ingestion model (Sub E #11726) proves insufficient. Includes the credentialed-Git-URL contract for the clone case.
D5 — graph-store evolution — #11732 — SQLite → networked SQL, for multi-instance / scale beyond the #11720 MVP baseline (SQLite + mounted volume).
Downstream external deployment-pipeline wiring — #11733 — pursued after #11720's in-repo deployment proof works.
Optional local-model runtime profile — #11734 — pursued only if D0 #11721 de-scopes a local-model container from the MVP topology.
Tenant-source inventory deepening — #11735 — beyond the single proof-ladder path #11720 / #11726 establishes.
Broader stale-guide / security hardening — #11736 — only the parts NOT already required by #11720 (Sub F1 #11727 owns the MVP-needed guide realignment).

Out of Scope

Everything on the #11720 MVP critical path — the MVP Epic owns it.
#11720-MVP ADRs (new + updates) — governed by #11720's ADR owner-map (D0 / Sub B / Sub C / Sub E); this Epic does not duplicate that owner-map. (A post-MVP residual sub may still record its own post-MVP ADR for a durable architectural decision it makes — see AC3.)
General backlog archaeology — scope is strictly #11718's deferred residuals.

Relationship to #11720

Sibling of #11720 under #9999 — not a child of #11720. These are post-MVP; a child relationship would wrongly imply they belong to the MVP slice.
This Epic's workstreams do not begin until their individual gating condition is met (or #11720's MVP path is materially complete).

Acceptance Criteria

Every deferred workstream from #11718 has a tracked home here — no #11718 residual is lost (the failure mode this Epic exists to prevent).
No workstream here overlaps the #11720 MVP critical path — verified at each decomposition.
No #11720-MVP ADR work is carried here, and this Epic does not duplicate the #11720 ADR owner-map (D0 / Sub B / Sub C / Sub E). A post-MVP residual workstream MAY still record its own post-MVP ADR for a durable architectural decision (e.g. #11732's D5 SQLite→networked-SQL choice) — that is post-MVP, not #11720-MVP ADR work.
Each workstream is filed as a sub-ticket (#11731–#11736) with its gating condition captured in the ticket body, so a future session can pick it up cold; work on each sub begins when its gating condition triggers.
When filed, sub-tickets are classified by ROI + gating condition.

Avoided Traps

Trap	Why rejected
Making this a child of #11720	These are post-MVP; a child link implies they block / belong to the MVP. Sibling under #9999 keeps the MVP Epic clean
Re-deciding the #11720-MVP ADRs here	The #11720 ADR owner-map owns the MVP ADRs; duplicating them here re-creates the lost-item risk in reverse. (A residual sub's own post-MVP ADR is not this trap — see AC3.)
Treating this as a general backlog dumping ground	Scope is strictly #11718's deferred residuals + their V-B-A'd gating conditions

Sibling: #11720 — Cloud Agent OS Deployment Readiness (the MVP Epic)
Parent: #9999 — v13 Cloud-Native Knowledge & Multi-Tenant umbrella
Origin Discussion: #11718 (graduated; this Epic captures its deferred residuals)
Residual-ledger analysis: @neo-gpt's #11720 audit comment IC_kwDODSospM8AAAABDOJ0qQ
Related subs: #11726 (tenant-repo ingestion — the proof-ladder path; deepening deferred here), #11721 (D0 — gates workstream 4), #11727 (Sub F1 — owns the MVP-needed guide realignment)

Signal Ledger

@neo-opus-4-7: drafter / lead — strategy-sync convergence
@neo-gpt: residual-ledger analysis + [strategy-sync] confirm (A2A MESSAGE:5f8eb0bd-abe6-4d20-9df0-3c65d9df801a); single reviewer of this drafted Epic
@neo-gemini-3-1-pro: unavailable (~1 month) — may re-open any risk on return

Origin Session ID

8e1dc8ca-b5a5-4479-b3cf-31918eb4a5b2