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.

Neo.mjs - Self-Evolving Software Organism

Frontmatter

title	>-
docs	Author Progressive Disclosure Skills guide and register cross-links
author	tobiu
state	Merged
createdAt	Apr 13, 2026, 9:12 PM
updatedAt	Apr 13, 2026, 9:14 PM
closedAt	Apr 13, 2026, 9:14 PM
mergedAt	Apr 13, 2026, 9:14 PM
branches	`dev` ← `agent/9986-progressive-disclosure-skills`
url	https://github.com/neomjs/neo/pull/9988

Merged

tobiu commented on Apr 13, 2026, 9:12 PM

Architecture Context

This PR introduces the final architectural guide in the Agent OS documentation series: Progressive Disclosure Skills (learn/agentos/ProgressiveDisclosureSkills.md).

Implementation Details

Docs the skill lifecycle triad (ticket-intake -> pull-request -> pr-review).
Expands on the token economics that mandate lazy-loading instructions.
Defines the progressive disclosure pattern and how skills compose dynamically with the root AGENTS.md mandates.
Registers the new guide inside the learn/tree.json structure.
Add reciprocal cross-links in SwarmIntelligence.md, StrategicWorkflows.md, and CodebaseOverview.md.

Resolves #9986

tobiu commented on Apr 13, 2026, 9:13 PM

Input from Antigravity (Gemini 3.1 Pro):

✦ # PR Review Summary

Status: Approved

Excellent job completing the Agent OS documentation suite. The Progressive Disclosure Skills guide is conceptually sound, clean, and perfectly articulates the lazy-loaded instruction strategy and token economics.

📊 Evaluation Metrics

[ARCH_ALIGNMENT]: 100 - Perfectly aligns with the Swarm Architecture and "Fat Ticket" documentation mandates.

[CONTENT_COMPLETENESS]: 100 - The structural cross-links and inventory were thoroughly mapped into the codebase trees.

[EXECUTION_QUALITY]: 100 - Clean markdown and comprehensive coverage of the problem space without hallucination.

[PRODUCTIVITY]: 100 - Fully resolved #9986.

[IMPACT]: 85 - High impact documentation piece for future human and agent orchestration understanding.

[COMPLEXITY]: 30 - Straightforward documentation and tree-linking endeavor.

[EFFORT_PROFILE]: Quick Win - High ROI for a relatively low-complexity integration task.

🕸️ Context & Graph Linking

Target Epic / Issue ID: Resolves #9986

Related Graph Nodes: AgentOS, SwarmIntelligence, StrategicWorkflows, CodebaseOverview

🧠 Graph Ingestion Notes

[KB_GAP]: None identified.

[TOOLING_GAP]: None identified.

[RETROSPECTIVE]: The progressive disclosure pattern is proving to be a highly effective mitigation for token exhaustion (n_ctx limit). Keeping discrete, atomic skill Markdown files heavily reduces noise in the primary system prompt loop while retaining complex tactical abilities.

📋 Required Actions

To proceed with merging, please address the following:

No actions required. Ready for merge!

Great work.