What is the Neural Link?

The Neural Link is a bi-directional bridge that connects AI agents directly to the Neo.mjs runtime. It lets agents inspect the Scene Graph, component state, event listeners, computed styles, and DOM rectangles, and mutate the running application in real time.

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

Neo.mjs maintains persistent application objects in a worker-backed Scene Graph instead of compiling application state away into ephemeral DOM nodes. That architecture enables multi-window orchestration, runtime permutation, and deep AI introspection.

What is Context Engineering?

Context Engineering shapes the information and tool environment around AI agents. Neo.mjs implements it through Knowledge Base, Memory Core, GitHub Workflow, and Neural Link MCP servers for frontier harnesses, plus a File System MCP server for internal Neo.ai.Agent local loops.

What is the Neo.mjs Agent OS?

The Neo.mjs Agent OS is the repository Brain: source code and services for Memory Core, Knowledge Base, Active Hybrid GraphRAG, DreamService, Golden Path synthesis, A2A coordination, and Neural Link tooling.

Frontmatter

id	15242
title	FM define-agent surface rebuild: AgentConfigCard + FleetSettingsPanel to the cockpit design SSOT (Lane D1 UX)
state	Open
labels	enhancementdesignai
assignees	[]
createdAt	1:15 PM
updatedAt	1:15 PM
githubUrl	https://github.com/neomjs/neo/issues/15242
author	neo-fable
commentsCount	0
parentIssue	14560
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
contentTrust
projected
quarantined	0
signals	[]
blockedBy	[]
blocking	[]

FM define-agent surface rebuild: AgentConfigCard + FleetSettingsPanel to the cockpit design SSOT (Lane D1 UX)

Open Backlog/active-chunk-6 enhancementdesignai

neo-fable commented on 1:15 PM

Context

Operator verdict on record (2026-07-16): the agents setup view "does not match the design specs in any way," and FM is "not functional yet, not even close" — design and UX are early priorities. This is the anchor leaf of the design-conformance + functionality wave for cornerstone 1 (v13.2).

The design SSOT is apps/agentos/design/fleet-manager-cockpit-plan.html (committed via #14512; @neo-opus-grace holds design authority on it). Its opening line defines this ticket's target by antithesis: "Mission control for a cross-model AI engineering team — not a settings form."

The Problem

The define-agent surface IS that settings form. It is the pre-design-sprint v0, functionally patched but never design-rebuilt:

apps/agentos/view/AgentConfigCard.mjs (219 lines), apps/agentos/view/FleetSettingsPanel.mjs (188 lines), apps/agentos/view/Accounts.mjs (609 lines) import zero fleet design primitives — no token layer, no fm-* classes, no StateDot / HealthSwatch / FamilyRail. Their last commits were data-plumbing only (#14998 sparse-config persistence, #14920 store de-singletonization).
The design sprint (D#14561 T-leaves) rebuilt view/fleet/** to the SSOT bar; the setup/admin surfaces were out of its slice and still carry the pre-sprint look the operator already ruled "uber brittle" (2026-07-04, on #14577's record).
SSOT Lane D1 states the functional half plainly: "Add-agent flow — username + PAT + harness type → the Node-side registry (#13031 exists); the UX for it does not."

The Architectural Reality

The Brain-side machinery is shipped: registry + credential store (#13031), control-plane facade (#13192), add-agent canonical readback round-trip (#14614). This leaf is UI composition, not new services.
The design system to compose FROM exists: token layer (#14578, TOKENS.md), AgentCard + card contract (#14598, CARD-CONTRACT.md), state/health/family primitives (#14593/#14635/#14637), both themes (#14681).
Credential boundary per #13058's matrix is non-negotiable: PAT/credential material never in browser persistent state, logs, URL, or visible echo; secret submission disabled when the Node-side bridge is absent; no fake success states (lifecycle-affordance rule).
apps/** gate applies: data-carrying UI binds data.Store of data.Model records; state.Provider at view roots; zero CSS-in-JS (SCSS token/skin layers only).

Acceptance Criteria

The define/add-agent surface is rebuilt from the fleet design primitives + token layer (light AND dark), replacing the form-generator look; AgentConfigCard / FleetSettingsPanel are each either rebuilt to the SSOT bar or retired and replaced under the cockpit shell — no second brittle surface left mounted.
The D1 flow works on the REAL surface: username + PAT + harness type → registry round-trip consuming #14614's canonical readback, with honest pending/degraded/failure states (no fake success; gated when the bridge is absent).
Credential boundary evidenced per #13058's matrix: test or audit proving no localStorage / sessionStorage / URL / log echo on submit.
Design sign-off by @neo-opus-grace (SSOT authority) on the shape before merge; cross-family review as usual.
NL-driven e2e witness of the add-agent journey on the rebuilt surface.

Out of Scope

The Accounts keeper-view / credential-management rebuild (D2-adjacent — its own wave leaf; #13521 context folds there) · live-wiring the cockpit stores off JSON seeds (C-lane wave) · wake/rate-limit telltales (#14537-adjacent) · catch-up view (#14620) · the drill regression (#15212, independent).

Epic #14560 (parent) · design SSOT via #14512 · #13058 (the v0 this replaces) · #14614 (readback consumed) · #13521 (adjacent, @neo-opus-vega) · #14577 (target-app decision + "uber brittle" verdict) · operator directive 2026-07-16 (design/UX first; ≥50-ticket wave).

Sweeps: latest-15 open checked 2026-07-16T11:12Z (latest #15241) — no equivalent; fleet-titled open census clean (7 open, none setup-surface); A2A herd-window clean (no in-flight claim on setup-surface work).

Origin Session ID: b42c5667-db78-4a0c-83a0-ce66c95ee0d9 Retrieval Hint: "FM define-agent setup surface rebuild design SSOT settings form D1"