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	10246
title	docs(ai): add fresh MCP client primitive technique to self-repair skill
state	Closed
labels	documentationaicoreneeds-re-triage
assignees	neo-gemini-pro
createdAt	Apr 23, 2026, 5:59 PM
updatedAt	3:19 PM
githubUrl	https://github.com/neomjs/neo/issues/10246
author	neo-gemini-pro
commentsCount	0
parentIssue	null
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
contentTrust
projected
quarantined	0
signals	[]
blockedBy	[]
blocking	[]
closedAt	Jun 6, 2026, 11:53 PM

docs(ai): add fresh MCP client primitive technique to self-repair skill

Closed v13.0.0/archive-v13-0-0-chunk-5 documentationaicoreneeds-re-triage

neo-gemini-pro commented on Apr 23, 2026, 5:59 PM

Context

During the #10139/#10241 Mailbox A2A integration effort, the agent successfully discovered a MailboxService.from tool-shape validation issue empirically via the unit-test + self-repair skills. This discovery was only possible because the agent used the ai/mcp/client primitive to spawn a fresh MCP server instance, bypassing the agent's own stale MCP cache.

The Problem

Agents operating on MCP servers they themselves are connected to often suffer from "ghost bugs"—where the agent's context or cached tool definitions do not match the live server state. The self-repair skill lacks explicit guidance on using a fresh MCP client to break out of this stale cache window.

The Architectural Reality

The ai/mcp/client infrastructure (e.g. mcp-cli.mjs) allows an agent to instantiate a completely clean, isolated client-server connection. This provides a "clean room" for diagnostics, isolated from the agent's primary long-lived host connection.

The Fix

Update the self-repair skill to explicitly document the ai/mcp/client primitive as a primary diagnostic tool for MCP server debugging.

Acceptance Criteria

Update .agent/skills/self-repair/SKILL.md with a new section documenting the "Fresh MCP Client Primitive".
Provide a concrete example in the skill of invoking ai/mcp/client/mcp-cli.mjs to test a tool shape and bypass stale cache windows.

Out of Scope

Refactoring the ai/mcp/client codebase itself.

Origin Session ID

Origin Session ID: 1c5933cc-a2d0-4296-ae3f-f4e815d385a2