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	13663
title	Handoff by-construction test-isolation: branch handoffFilePath so test runs don''t clobber the tracked sandman_handoff.md (#12435 slice)
state	Closed
labels	aitestingarchitecturemodel-experience
assignees	neo-opus-grace
createdAt	Jun 20, 2026, 10:44 PM
updatedAt	3:20 PM
githubUrl	https://github.com/neomjs/neo/issues/13663
author	neo-opus-grace
commentsCount	1
parentIssue	null
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
contentTrust
projected
quarantined	0
signals	[]
blockedBy	[]
blocking	[]
closedAt	Jun 21, 2026, 11:31 AM

Handoff by-construction test-isolation: branch handoffFilePath so test runs don't clobber the tracked sandman_handoff.md (#12435 slice)

Closed v13.1.0/archive-v13-1-0-chunk-4 aitestingarchitecturemodel-experience

neo-opus-grace commented on Jun 20, 2026, 10:44 PM

Context

Ada's #12435 surface map (A2A) flagged handoffFilePath (ai/mcp/server/memory-core/config.template.mjs:392) as a live-store-write vector with no test/prod branch: GoldenPathSynthesizer.mjs:1296 reads aiConfig.handoffFilePath and writeFileSyncs the handoff — so an offline GoldenPathSynthesizer run under a test harness clobbers the TRACKED resources/content/sandman_handoff.md.

Unlike the trajectories vector (where backup.mjs:104 hardcodes the path), the handoff writer already reads the config leaf — so this is a clean by-construction fix: just branch the leaf.

The fix

Mirror the collections.memory / storagePaths.graph by-construction shape: split handoffFilePath → handoffFilePathProd (the tracked resource, unchanged) + handoffFilePathTest (a disposable .neo-ai-data/test/ path), resolved via a formulas entry gated on the canonical storagePaths.useTestDatabase (UNIT_TEST_MODE) toggle. Under UNIT_TEST_MODE, aiConfig.handoffFilePath resolves to the test path by construction → no singleton mutation, and GoldenPathSynthesizer never clobbers the tracked handoff.

AC

aiConfig.handoffFilePath resolves to the tracked path in prod, the disposable test path under UNIT_TEST_MODE.
Prod path byte-unchanged (resources/content/sandman_handoff.md).
A unit test asserts the UNIT_TEST_MODE resolution.
Edit config.template.mjs (the tracked SSOT) only; config.mjs re-materializes via --migrate-config.

Out of scope

The trajectories + concepts vectors (separate; trajectories needs hardcode-removal per my A2A); the graph-spec construct-config-injection primitive.

Refs #12435, #13658, #13624.