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	13297
title	Wire neural-link HealthService to RuntimeFreshnessService (freshness parity)
state	Closed
labels	enhancementaiarchitecturemodel-experience
assignees	neo-opus-grace
createdAt	Jun 15, 2026, 5:59 AM
updatedAt	Jun 21, 2026, 2:17 AM
githubUrl	https://github.com/neomjs/neo/issues/13297
author	neo-opus-grace
commentsCount	0
parentIssue	13289
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
blockedBy	[]
blocking	[]
closedAt	Jun 15, 2026, 9:18 AM

Wire neural-link HealthService to RuntimeFreshnessService (freshness parity)

Closed v13.1.0/archive-v13-1-0-chunk-3 enhancementaiarchitecturemodel-experience

neo-opus-grace commented on Jun 15, 2026, 5:59 AM

Context

Sub-leaf of #13289. #13289 found (via the #13286 stale-bridge incident) that the neural-link MCP bridge has zero runtime-freshness tracking, while Memory Core / Knowledge Base / GitHub Workflow each surface a runtimeFreshness block (digest-based, cloud-safe, via the shared RuntimeFreshnessService). This leaf delivers the concrete parity wiring; #13289 stays open for its broader goal — cheap source-code staleness detection (the #13286 class proper), which is an open design question (Ideation Sandbox candidate).

The Problem

ai/services/neural-link/HealthService.mjs returns a bridge-connection status payload but no freshness diagnostic. So a long-lived neural-link bridge process can be bridge-healthy while its config/OpenAPI schema has moved underneath it, with no signal — unlike the other three MCP servers.

The Architectural Reality

ai/mcp/server/shared/services/RuntimeFreshnessService.mjs already provides a per-consumer createTracker({files, serviceName, ...}) digesting configDigest (ai/config.mjs) + openApiDigest (the server's openapi.yaml), and ai/services/{memory-core,knowledge-base,github-workflow}/HealthService.mjs are the existing consumers. neural-link is the only MCP server not wired, and it has both required inputs (ai/config.mjs + ai/mcp/server/neural-link/openapi.yaml).

The Fix

Mirror the KB consumer pattern in neural-link's HealthService: a module-level RuntimeFreshnessService.createTracker(...) (config + OpenAPI digests, serviceName: 'Neural Link MCP server'), the #runtimeFreshnessTracker field + the boot-identity / reader / cache accessors, a resolveRuntimeFreshness() method + clearCache(), and runtimeFreshness: await this.resolveRuntimeFreshness() in the healthcheck payload.

Acceptance Criteria

neural-link's healthcheck payload includes a runtimeFreshness block with status (current/stale/unknown), via the shared RuntimeFreshnessService.
Git-independent (no gitHead; digest-only, cloud-safe — neural-link supplies no rootDir).
Unit-verified: injected matching identity → current; changed config/OpenAPI digest → stale with Neural-Link restart guidance; unreadable identity → unknown without throwing; gitHead drift ignored.

Out of Scope

Source-code staleness detection (the #13286 class proper) — the harder, cheap-mechanism-TBD design question; stays on #13289 (Ideation Sandbox candidate).

Decision Record impact

none. Aligned-with the existing RuntimeFreshnessService + the #12778 gitHead-removal (cloud-safe digest design).

Parent: #13289 (umbrella — stays open for code-staleness). Surfacing incident: #13286. Shared infra: ai/mcp/server/shared/services/RuntimeFreshnessService.mjs. Aligned anchor: #12778.

Live latest-open sweep: checked latest 20 open issues at 2026-06-15T03:58Z; no equivalent found.

Release classification: post-release (Agent Harness infra reliability) — boardless.

Origin Session ID: 0f5d9f1d-0683-452d-aac1-f467297186ac

Authored by Claude Opus 4.8 (Claude Code), @neo-opus-grace (Grace).