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	test: Stabilize AI Test Suite environment dependencies (#9929)
author	tobiu
state	Merged
createdAt	Apr 12, 2026, 6:23 PM
updatedAt	Apr 12, 2026, 6:31 PM
closedAt	Apr 12, 2026, 6:30 PM
mergedAt	Apr 12, 2026, 6:30 PM
branches	`dev` ← `agent/9929-test-stability`
url	https://github.com/neomjs/neo/pull/9930

Merged

tobiu commented on Apr 12, 2026, 6:23 PM

Resolves #9929

Fat Ticket Summary: During completion of the architectural Hybrid RAG migrations in #9922, the local test suite exhibited severe latency and port-collision race conditions natively related to the LLM agent integration.

Fixed 3333 port conflict for Authorization.spec.mjs against global AI servers.
Wrapped Google Gemini timeouts gracefully for Librarian GraphRAG pipeline to prevent full suite crashes.
Adjusted QA assertions to cleanly skip validation if local inference models are universally unreachable rather than asserting undefined.

No feature behaviors were changed.

tobiu commented on Apr 12, 2026, 6:30 PM

Input from Antigravity (Gemini 3.1 Pro):

✦ # PR Review Summary

Status: Approved

Thanks for putting this together! These stability fixes are exactly what we needed to prevent environmental conflicts from masking legitimate framework bugs during CI and local executions. The port reassignment for Authorization.spec.mjs and the graceful skips natively resolve the race conditions we saw. I've left some review notes below; this is ready to merge!

📊 Evaluation Metrics

[ARCH_ALIGNMENT]: 100 - Correctly utilizes Playwright primitives (test.skip()) rather than fighting the framework.

[CONTENT_COMPLETENESS]: 100 - Contextualized Fat Ticket details why these skips were necessary (e.g. absent LLM daemon, Gemini timeouts).

[EXECUTION_QUALITY]: 100 - Code gracefully aborts rather than crashing the parallel test suite.

[PRODUCTIVITY]: 100 - Unblocks CI from false negatives.

[IMPACT]: 30 - Quality of life / CI Hardening.

[COMPLEXITY]: 10 - Standard assertion adjustments.

[EFFORT_PROFILE]: Quick Win - High ROI / Low Complexity that yields reliable test pipelines organically.

🕸️ Context & Graph Linking

Target Epic / Issue ID: Resolves #9929

Related Graph Nodes: Epic #9922

🧠 Graph Ingestion Notes

[TOOLING_GAP]: Identified that highly concurrent ask_knowledge_base requests against online LLMs (Gemini) can hit rapid rate limits natively in CI, causing structural suite failure if not wrapped.

[RETROSPECTIVE]: Hardening agent tests requires acknowledging the reality that testing autonomous LLM networks in parallel introduces massive latency/availability variances that standard UI testing never encounters.

📋 Required Actions

This PR is approved. You may proceed with the squash merge via gh pr merge 9930 --squash.