Elevate Dream Pipeline in v13 release notes

Context

The v13 release-note work is moving from small cut-lines toward the release's real headline weight. The operator explicitly challenged the swarm not to over-prioritize final polish while the release contains a massive v13 workstream. Local release appendix evidence currently reports 1,228 merged PRs and 1,587 closed issues since the v12.1 boundary, and the count must be refreshed at release cut because the live window is still moving.

This ticket exists because learn/agentos/DreamPipeline.md is not a polishing detail. It is a v13-native mechanism document for Neo's Brain: DreamService, REM digestion, Native Edge Graph extraction, Golden Path synthesis, and resources/content/sandman_handoff.md as the generated advisory handoff. The release notes should make this legible as a major shipped capability, not bury it as one anchor in a broad Brain paragraph.

Pre-creation evidence on 2026-06-08:

• Live latest-open sweep: checked latest 30 open issues with number, title, author, labels, and URL. Found parent epic #12696 and broad feed ticket #12698, but no dedicated Dream Pipeline / Golden Path release-note sub.
• Live exact duplicate search for Dream Pipeline Golden Path release notes found only #10120, which is an older Golden Path product/visibility issue, not a v13 release-note content sub.
• Local exact sweep across resources/content/issues and resources/content/discussions found Discussion https://github.com/orgs/neomjs/discussions/12694 and older identity/GraphRAG/Dream architecture discussions, but no equivalent open issue for this release-note chapter.
• Semantic KB ticket sweep for ticket v13 release notes Dream Pipeline Golden Path Native Edge Graph REM sandman_handoff surfaced v13/Dream architecture tickets such as #12066 and #11964, not a duplicate release-note content leaf.
• Source freshness check: git ls-tree shows learn/agentos/DreamPipeline.md, ai/graph, and ai/services/graph absent from origin/main and present on origin/dev. The current tree has 18 files across learn/agentos/DreamPipeline.md, ai/graph, ai/services/graph, and resources/content/sandman_handoff.md; git diff --shortstat origin/main...origin/dev -- ai learn/agentos/DreamPipeline.md resources/content/sandman_handoff.md reports 347 files changed and 94,278 insertions.

The Problem

The current v13 release-note artifact already names Neo as a self-evolving software organism, but the Dream Pipeline / Golden Path mechanism can still read as a supporting proof point instead of a release-defining capability. That underweights one of the biggest v13 deltas: Neo now has a native graph-backed feedback loop that digests agent work, infers topology, detects gaps, and forecasts high-ROI next work.

Without a dedicated content pass, the final artifact risks over-representing the most recent polish and under-representing the new Brain substrate that makes the release qualitatively different from v12.1.

The Architectural Reality

learn/agentos/DreamPipeline.md defines DreamService as the forecasting engine that runs offline or at startup, digests raw session memories into graph intelligence, and synthesizes a Golden Path. The document includes two release-ready diagrams: the six-phase REM pipeline and the closed loop from sessions to DreamService to Native Edge Graph to sandman_handoff.md back to peer maintainers.

Key public source anchors:

• learn/agentos/DreamPipeline.md documents Phase 0 workspace ingestion, Phase 1 tri-vector extraction, Phase 2 topological conflict detection, Phase 3 deterministic capability-gap inference, Phase 4 graph/vector apoptosis, and Phase 5 Golden Path synthesis.
• ai/services/graph/GoldenPathSynthesizer.mjs describes Golden Path synthesis as Hybrid GraphRAG over vector similarity plus graph topology, then writes resources/content/sandman_handoff.md.
• resources/content/sandman_handoff.md is a current generated output surface from the Native Edge Graph audit.
• README.md already frames Native Edge Graph, DreamService, and Golden Path as part of the Brain pillar.

The Fix

Update the v13 release-note artifact so Dream Pipeline / Golden Path becomes a major Brain chapter or clearly elevated subsection. The prose should make the mechanism legible:

• sessions become graph intelligence;
• graph intelligence becomes Golden Path math;
• Golden Path math produces an advisory handoff;
• peer maintainers still self-select work rather than becoming workers for an auto-assignment queue;
• the loop changes future work because completed work changes the graph.

The release note may reuse a Mermaid diagram if it materially clarifies the loop, but the diagram must support the thesis rather than decorate the page.

Contract Ledger Matrix

Decision Record impact

None. This is release-content synthesis over already-committed v13 substrate; it does not amend an ADR.

Discussion Criteria Mapping

Acceptance Criteria

• AC1: resources/content/release-notes/v13.0.0.md contains a major Dream Pipeline / Golden Path Brain passage, either as a dedicated chapter or clearly elevated subsection.
• AC2: The passage cites public source anchors for learn/agentos/DreamPipeline.md, ai/services/graph/GoldenPathSynthesizer.mjs, and resources/content/sandman_handoff.md or their PR/issue equivalents.
• AC3: The prose explains the actual loop: sessions -> Native Edge Graph intelligence -> Golden Path math -> advisory handoff -> peer-selected work -> new sessions.
• AC4: The release note avoids over-claiming Golden Path as automatic assignment. It must preserve the flat peer-team / advisory forecast boundary.
• AC5: The prose keeps v13's scale visible without pasting the exhaustive PR/issue corpus into the narrative body; final counts remain governed by the release-cut refresh checklist.
• AC6: Public-surface hygiene is checked: no client names, no private deployment specifics, no unverifiable market claims.

Out of Scope

• Final narrative-arc polish for the whole release note.
• Changing DreamService, GoldenPathSynthesizer, Native Edge Graph, or Memory Core runtime code.
• Generating the exhaustive release appendix.
• Editing the future release-notes skill #10321.
• Claiming that #12065 or any other open Dream/REM epic is fully resolved.

Avoided Traps

• Rejected: treating this as a final polish pass. The task is content/context weighting for a major v13 Brain capability.
• Rejected: turning Golden Path into a work-assignment hierarchy. The source says advisory forecast, and AGENTS.md requires flat peer-team agency.
• Rejected: raw changelog dumping. The release body should explain the mechanism and link evidence; exhaustive history belongs in appendices.
• Rejected: claiming all Dream/REM work is complete. This ticket is about the release-note chapter for shipped substrate and source-backed capability, not closing every REM backlog item.

Related

• Parent epic: #12696
• Current release-note feed: #12698
• Release-note narrative discussion: https://github.com/orgs/neomjs/discussions/12694
• Future release-notes skill: #10321
• Dream / REM architecture context: #12065

Origin Session ID: e8f07ef9-ef7e-4815-8ff4-7abe13720621

Retrieval Hint: v13 release notes Dream Pipeline Golden Path Native Edge Graph Hybrid GraphRAG sandman_handoff major Brain chapter operator 1300 PR scale