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	fix(ai): degrade Golden Path interpretation explicitly (#13983)
author	neo-gpt
state	Merged
createdAt	Jun 25, 2026, 4:26 AM
updatedAt	Jun 25, 2026, 4:25 PM
closedAt	Jun 25, 2026, 4:25 PM
mergedAt	Jun 25, 2026, 4:25 PM
branches	`dev` ← `codex/13983-golden-path-degraded-interpretation`
url	https://github.com/neomjs/neo/pull/13988
contentTrust
projected
quarantined	0
signals	[]

Merged

neo-gpt commented on Jun 25, 2026, 4:26 AM

Resolves #13983

Golden Path handoff generation is narrowed to the corrected contract: the generated ## Computed Golden Path (Strategic Recommendation) section now gets its own Captured at: YYYY-MM-DD HH:mm UTC row, and invalid/failed Strategic Interpretation generation renders an explicit degraded reason instead of synthetic priority prose.

The successful computed route remains the expected happy path: section headline, Captured at directly below the headline, the Tri-Vector recommendation intro, numbered issue-* / discussion-* recommendations, and > **Strategic Interpretation:**.

This PR does not emit a Golden Path Forecast Status section, route-provenance paragraph, or candidate-count diagnostics inside the successful computed route. Those mechanics are already represented by the existing Computed Golden Path section and DreamPipeline.md; duplicating them inside the handoff was rejected as bloat.

Evidence: L2 focused static/unit verification completed locally after rebasing onto current origin/dev. Residual: GitHub CI is running on the latest pushed head; L3 post-merge orchestrator validation must inspect the next generated resources/content/sandman_handoff.md.

Deltas from ticket

The ticket body allowed compact deterministic fallback; the owner comment on #13983 rejected that path. Follow-up operator corrections also rejected the forecast-status metadata block and the document-title freshness row. The branch now keeps only:

Computed Golden Path section-local Captured at timestamp
explicit non-synthetic degraded reason when the model-generated Strategic Interpretation is unavailable
exact happy-path test coverage for mixed issue/discussion computed recommendations

The branch was rebased after #13987 landed and preserves #13987 Active PR Cycle freshness/degraded rendering without adding new Golden Path forecast bloat.

Test Evidence

node --check ai/services/graph/GoldenPathSynthesizer.mjs
node --check test/playwright/unit/ai/services/graph/GoldenPathSynthesizer.spec.mjs
git diff --check origin/dev...HEAD
node ./buildScripts/util/check-block-alignment.mjs --staged
npm run agent-preflight -- test/playwright/unit/ai/services/graph/GoldenPathSynthesizer.spec.mjs
npm run test-unit -- test/playwright/unit/ai/services/graph/GoldenPathSynthesizer.spec.mjs -> 36 passed

Post-Merge Validation

Pull dev and restart the orchestrator.
Confirm the next generated resources/content/sandman_handoff.md Computed Golden Path section contains Captured at: ... UTC.
Confirm successful computed recommendations keep numbered issue-* / discussion-* rows followed by > **Strategic Interpretation:**.
Confirm the generated handoff does not contain ## Golden Path Forecast Status.
Confirm invalid/offline Strategic Interpretation generation renders a degraded reason, not deterministic fallback rationale.

Commits

0d41023d0e - fix(ai): degrade Golden Path interpretation explicitly (#13983)
092c55c732 - test(ai): cover Golden Path happy path format (#13983)

Authored by Euclid (GPT-5, Codex Desktop). Session 019ef378-527d-7393-bc74-ec3a1d3f2ddf.

neo-gpt commented on Jun 25, 2026, 5:47 AM

Merge-order note from preflight:

#13987 and #13988 both touch ai/services/graph/GoldenPathSynthesizer.mjs at the helper-method seam and conflict in both merge orders.
The conflict is additive, not conceptual: #13987 adds Active PR Cycle freshness/degraded rendering; this PR adds explicit Golden Path forecast/degraded strategic-interpretation status.
Recommended sequence after #13986/#13990: merge #13987 first, then rebase this PR on refreshed dev and resolve the helper-seam conflict.
I did not stack/rebase this PR now because that would make the diff non-clean against dev before the first sibling PR lands.

Validation already run for the preflight: git merge-tree in both orders confirmed the conflict; public KB/MC smoke is unchanged until #13986/#13990 are deployed.

neo-gpt commented on Jun 25, 2026, 6:34 AM

Merge-order note after stack preflight: please do not merge this before #13987. #13987 owns the stronger correctness gate (wroteHandoff scheduler enforcement + renderActivePrCycleState). This PR should be rebased/updated after #13987 lands so it preserves that write-proof path and only layers the Strategic Interpretation degraded reason / forecast-status rendering on top.

tobiu APPROVED reviewed on Jun 25, 2026, 4:23 PM

No review body provided.