Rebuild v13 release note around reader-value arc

Context

The v13 release-note artifact exists, but the current document still reads more like an evidence-rich planning memo than the major-release story established by the local exemplars. The operator reset the lane on 2026-06-08 after multiple failed iterations: the release note must make a CTO, CEO, architect, or LLM immediately understand:

1. what v13 is about,
2. why they should care,
3. what is in it for them.

The cited bar is resources/content/release-notes/chunk-2/v12.1.0.md, v12.0.0.md, and v11.24.0.md: named thesis, origin conflict, industry contrast, prioritized hero cases, war stories, and a grouped changelog/proof section. The current resources/content/release-notes/v13.0.0.md has valuable material, but it still starts with process/accounting, has a How To Read v13 section, includes too many representative anchors in the narrative body, and gives supporting beats such as GitLab provider work and Grid polish more weight than the v13 epoch shift.

This ticket is the fresh full-document rewrite authority after the rejected front-slice / voice-first path.

V-B-A evidence before filing:

• Live latest-open sweep on 2026-06-08 checked 60 open issues with number, title, author, labels, assignees, and URL. Related open issues found: #12696 parent epic, #12782 cost-safety chapter, and #12786 Sandman generated-handoff beat. No equivalent whole-document reader-value rewrite leaf was open.
• Exact live search v13 release notes narrative arc returned #12696, #12782, and #12786 only.
• Exact live all-state search v13 release notes rewrite found closed/related release-note leaves (#12695, #12699, #12702, #12723, #12733, #12783) plus #12786; the closest prior hero-piece path #12783 is closed after operator rejection and is not a live authority to continue.
• Local exact sweep across resources/content/issues and resources/content/discussions for v13 release notes, whole-document, How To Read v13, goosebumps, reader contract, and narrative arc found release-content infrastructure and old narrative-arc tickets, but no equivalent active rewrite leaf.
• Semantic ticket sweep with query_documents(type='ticket') for v13 release notes whole document narrative arc rewrite reader contract did not surface an equivalent ticket.
• ask_knowledge_base(type='ticket') for the same duplicate shape returned no matching narrative-arc rewrite ticket.
• Source inspection of the current release note confirmed the problematic surfaces: TL;DR/count-heavy opener, ## How To Read v13, ai/graph ... absent from main style creator-story leakage, dense representative anchors inside narrative sections, and large Body/Grid / GitLab supporting sections.
• Source inspection of the v12/v11 exemplars confirmed the target structure: named release identity first, origin/conflict next, prioritized hero case studies, war stories with proof, then grouped changelog evidence.

The Problem

The release note currently explains the v13 substrate to insiders; it does not yet make an outside reader want the thing.

v13 is not primarily:

• a list of PR counts,
• a file-count proof table,
• a GitLab MCP provider milestone,
• a Grid-polish release,
• or a release-note reading guide.

v13 is the release where a world-class multi-threaded application engine gained the operating substrate it needed to maintain itself: a Brain, a named cross-family AI maintainer institution, Dream / Golden Path, and multi-tenant deployment so that the same Agent OS can be pointed at another codebase. The current document contains many of those facts, but the ordering and narrative weight obscure the big picture.

The Architectural Reality

Release notes live at resources/content/release-notes/ and are public-facing source material for the portal, SEO surfaces, KB ingestion, LLM readers, and human release evaluation.

The local precedent matters more than a generic changelog template:

• v12.0.0.md is a major-release exemplar: named thesis (Zero Overhead), origin story (The Crucible), industry contrast, prioritized case studies, diagrams/code/proof, then changelog.
• v12.1.0.md centers the hero feature (TreeStore / TreeGrid) and uses supporting phases rather than a flat feature inventory.
• v11.24.0.md and v11.23.0.md tell concrete paradox/incident stories before listing the implementation corpus.

The v13 reader contract should be explicit in the prose shape even if no How To Read section remains:

The Fix

Rewrite resources/content/release-notes/v13.0.0.md around the full-document reader-value arc, reusing existing verified content where it serves the story and demoting over-weighted supporting material into concise secondary beats or the grouped changelog.

Required narrative shape:

1. Open with a named v13 thesis that earns attention immediately. Candidate direction: Neo Learned to Maintain Itself, sharpened as the engine growing a Brain, an Institution, and a Dream that can be deployed onto your codebase.
2. Replace process/accounting-first structure with an origin story: v12.1 had a world-class Body plus amnesic assistance; v13 built durable memory, graph, peer-review institution, Dream, and cloud deployment because maintaining a 23,000+ commit engine at AI velocity requires more than one stateless agent.
3. State why the reader should care: the industry is bolting amnesic models onto workflows, while Neo proves a persistent, cross-lab AI engineering institution maintaining a real engine in public.
4. State what is in it for them: a deployable Agent OS for their codebase, durable context beyond one window, graph-backed prioritization, cross-family review, and a live application engine agents can inhabit.
5. Prioritize hero chapters: Brain, Swarm / Institution, Dream, and Cloud Agent OS. Body/Grid, GitLab, docs, provider work, and UI polish support those heroes unless they directly carry one.
6. Move most PR anchors out of the narrative body and into the grouped changelog / appendix proof areas. Keep only rare hero anchors where the story depends on them.
7. Keep counts as final-cut facts only. Do not churn PR/issue counts during drafting.
8. Preserve honesty boundaries: Body/Grid is not fully finished; Dream/Golden Path is advisory; generated Sandman handoff evidence includes residual graph-audit caveats; public prose does not mention private deployment vehicle specifics.

Contract Ledger Matrix

Decision Record impact

none. This is public release-note restructuring within existing content architecture. It does not amend ADRs or change runtime behavior.

Acceptance Criteria

• resources/content/release-notes/v13.0.0.md no longer contains a How To Read v13 section.
• The opening answers the reader contract without meta-instructions: what v13 is about, why it matters, and what the reader gets from it.
• The document has a clear named thesis / release identity comparable in narrative role to the v12/v11 exemplars.
• The primary narrative hierarchy is Brain / Swarm-Institution / Dream / Cloud Agent OS, with Body/Grid, GitLab provider work, docs, and UI polish demoted to supporting beats unless directly connected to the hero arc.
• PR / issue anchors are materially reduced in the narrative body and concentrated in grouped changelog / appendix-style proof areas.
• The prose avoids absent from main / inventory-first creator-story phrasing in the hero narrative.
• Counts are not refreshed as part of this draft unless the release is at final cut; any existing count wording is either removed from the opener or framed as final-cut-refresh-needed.
• Body/Grid guardrails remain intact and do not claim completion of #9491 / #9872 / #9492 / #12758 ahead of live evidence.
• Dream / Sandman generated-handoff evidence is folded into the Dream hero or digest with its honest residual caveat from #12786.
• No .agents/skills/**, release-workflow skill, runtime source, or release-publish script is modified by this ticket.

Out of Scope

• Final release-count refresh.
• Publishing v13, merging release PRs, or executing release-line work.
• Creating or modifying the future release-notes skill #10321.
• Runtime changes to Memory Core, Native Edge Graph, DreamService, GitHub Workflow, GitLab Workflow, or Grid.
• Full blog/social announcement copy.
• Private deployment vehicle specifics.

Avoided Traps

• Reusing #12786 for the whole rewrite. It is intentionally a narrow Sandman handoff evidence beat.
• Continuing #12783 / #12785 as if the rejected voice-first path were still valid.
• Treating #12695 as sufficient authority after it was closed Done; the artifact exists, but the current correction needs a fresh, narrow leaf.
• Making GitLab provider work or Grid polishing read as v13 headline material.
• Proving every paragraph inline with random PR anchors. The body tells the story; the grouped changelog carries proof.
• Asking the reader how to read the release note. If the document needs that, the structure has failed.

Related

Parent epic: #12696 Related: #12695, #12698, #12699, #12702, #12716, #12723, #12733, #12777, #12782, #12783, #12786, #10321

Origin Session ID: 720fc09a-e8ff-467d-b039-28120312964d

Retrieval Hint: v13 release note reader contract goosebumps whole-document narrative arc Brain Swarm Dream Cloud How To Read v13 operator reset Vega compaction