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	7946
title	Feat: Implement Client-Side Tool Validation in MCP Client
state	Closed
labels	enhancementai
assignees	tobiu
createdAt	Nov 30, 2025, 11:59 AM
updatedAt	Nov 30, 2025, 2:40 PM
githubUrl	https://github.com/neomjs/neo/issues/7946
author	tobiu
commentsCount	1
parentIssue	7931
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
blockedBy	[]
blocking	[]
closedAt	Nov 30, 2025, 2:40 PM

Feat: Implement Client-Side Tool Validation in MCP Client

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

Closed v11.14.0 enhancementai

tobiu commented on Nov 30, 2025, 11:59 AM

Goal

Implement client-side validation for tool calls in Neo.ai.mcp.client.Client.

Context

Our MCP servers already validate tool shapes using Open API specs (ai/mcp/server/toolService.mjs). However, the current Neo.ai.mcp.client.Client implementation sends raw arguments to the server without pre-validation.

Requirement

The Client should:

Fetch the tool schema via listTools().
Before sending a request via callTool(), validate the arguments against the schema.
Throw a descriptive error if validation fails, providing immediate feedback to the agent (the caller) about the contract violation.

Value

Safety: Prevents "derailment" caused by sending malformed data to the server.
Feedback: Gives agents precise details on what they did wrong (contract violation) without needing a round-trip to the server.

tobiu added the enhancement label on Nov 30, 2025, 11:59 AM

tobiu added the ai label on Nov 30, 2025, 11:59 AM

tobiu added parent issue #7931 on Nov 30, 2025, 12:00 PM

tobiu assigned to @tobiu on Nov 30, 2025, 12:03 PM

tobiu cross-referenced by #7947 on Nov 30, 2025, 1:24 PM

tobiu referenced in commit 79dd59f - "Feat: Implement Client-Side Tool Validation in MCP Client #7946" on Nov 30, 2025, 2:38 PM

tobiu Nov 30, 2025, 2:40 PM

Input from Gemini 2.5 pro:

✦ I have successfully implemented client-side tool validation in Neo.ai.mcp.client.Client.

Verification of Necessity

I confirmed that the @modelcontextprotocol/sdk client does not perform schema validation on tool arguments before sending requests. It only validates the JSON-RPC protocol structure. This was proven by disabling my custom logic, which resulted in the server (not the client) catching the invalid input.

Implementation

Schema Caching: Client.mjs now caches tool schemas during initialization.

Validation: A new validateSchema method enforces the contract (types, required fields, enums) on the client side.

Interception: callTool now validates arguments before the request leaves the client.

This enhancement ensures immediate feedback for agents and prevents unnecessary network round-trips for malformed requests.

tobiu closed this issue on Nov 30, 2025, 2:40 PM

tobiu cross-referenced by #7949 on Nov 30, 2025, 2:50 PM