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	>-
author	neo-opus-ada
state	Merged
createdAt	May 30, 2026, 2:24 AM
updatedAt	May 30, 2026, 2:36 AM
closedAt	May 30, 2026, 2:36 AM
mergedAt	May 30, 2026, 2:36 AM
branches	`dev` ← `agent/12106-services-mcp-codemod`
url	https://github.com/neomjs/neo/pull/12185

Merged

neo-opus-ada commented on May 30, 2026, 2:24 AM

Summary

Stage 4 of Epic #12101 — consumer-side codemod migrating ai/services + ai/mcp off the non-getter aiConfig anti-patterns (Discussion #12100 sub-classes A / B / D-prod / E) onto the canonical Proxy-routed read + the no-hidden-default contract. Behaviour-preserving.

Deltas

A (12): aiConfig.data.X → aiConfig.X (drop the Proxy bypass) — DreamService, ConceptDiscoveryService, GapInferenceEngine, DatabaseService, SessionService, GraphService.
D-prod (8): bootstrap aiConfig.data.debug = true → aiConfig.debug = true (proxy set-trap → setData) — the 6 mcp-servers + run-bridge + mcp-cli. (No aiConfig.set() method exists — Discussion #12100 hypothesized one; the proxy assignment is the write path.)
B (3): ProviderReadinessHelper default-params config = aiConfig.data → config = aiConfig.
E (scalar + tenant): removed config-backed redundant ?? <literal> fallbacks — guideGapWeightThreshold/minSourceLength/prScanLimit, mcpSyncMaxChunks (×2), defaultTenantId/defaultRepoSlug (DocumentService/QueryService/SearchService). The config leaf provides each value verbatim; behaviour-identical.

Test Evidence

963 ai/services unit specs pass. The 8 failing memory-core specs are pre-existing env-flaky (need live ollama/models/MC) — confirmed failing identically on stashed-clean dev (git stash → same 8 fail → pop). node --check clean on all 19 files.

Evidence: every removed E default is config-backed — verified the owning leaf(...) in the templates (guideGapWeightThreshold: leaf(0.8), mcpSyncMaxChunks: leaf(50), defaultTenantId: leaf('neo-shared'), …). The tenant filters now read aiConfig.defaultTenantId verbatim (= 'neo-shared' for any current config), so the fail-closed isolation is unchanged.

Post-Merge Validation

None for the bulk — behaviour-preserving.

⚠️ Open sub-decision (deliberately NOT in this PR): the 9 KB Source ?? '<path>' fallbacks are #11660's deliberate byte-equivalence-for-stale-configs shims. The sourcePaths leaf defines all 9 verbatim, so the no-hidden-default contract says remove them — but removal breaks KB ingestion for operator overlays predating #11660's sourcePaths block. Surfaced as a config-SSOT-vs-backward-compat call before removal.

FAIR-band: n/a — operator-directed epic lane (#12101 Stage 4).

Refs #12106.

Authored by Claude Opus 4.8 (Claude Code, 1M context).