Terminal predicate: .github/RELEASE_NOTES/v13.3.0.md tells what 13.3 shipped to the release-notes bar and passes the §7 cut-readiness checklist at the cut head, so the #19652 cut publishes it.
Problem
13.3 has no release notes and no epic for them. publish.mjs prepare stops on .github/RELEASE_NOTES/v13.3.0.md is missing (buildScripts/release/releaseState.mjs), and the cut record #19652 leaves the notes to this epic (release-notes workflow §6). The cut cannot start until the note exists.
The window is short and fast. 13.2.0 was published at 2026-10-11T22:37Z, the ROADMAP (#19642) sets a one-to-two-week cadence, and milestone v13.3 is due 2026-10-25. By 2026-10-13T14:02Z, GitHub search counted 34 merged PRs in neo, 11 in neo-stage, 6 in Brain and 0 in Institution. These are search counts; the note's scope comes from the corpus run at the cut (§2). The arcs are hours old, so mining them costs less now than it ever will again (#14800's rationale).
The bar includes pictures. The operator's verdict on 13.2's iteration-6 draft was "not a single screenshot, not compelling" (#14800), and the published 13.2.0 note carries nine images.
Why an epic: the workflow (§1) makes release notes an epic by design. Iteration 1 merges early behind a banner, mining-driven leaves refine one arc each, and the last leaf passes §7. Each is a one-PR leaf with its own cross-family review.
Intended solution
- The three cornerstones lead, each mined before it is drafted (§3): content-sized rows (#19658, from D#19645, closed; steward @neo-opus-ada), header menus (#19643, from D#19646; steward @neo-opus-vega) and a cut the maintainers can run (#19637, closed, with #19652 as its proof; steward @neo-opus-grace). Each steward deposits one mining comment here once their arc has landed: Memory Core sessions, PR chain, timeline, numbers. Authors assemble from the deposits instead of re-mining cold.
- Screenshots of what the gate names: the 10k wrapped-text grid holding its rows through scrolling, and a header menu filtering by checked values with live counts.
- Also in 13.3, taken from the §2 scope run rather than memory. Known candidates: the Brain-gated e2e tier on every Engine PR (#19657, after #19656's
dev-push burn-in), the Portal's featured Workstation reaching the package (#19640), and the launch running beside the gate (#14790).
- Honest bounds: e2e-proven behaviour stays apart from bounded witnesses, such as the header menu's 100k timing (#19711: "a bounded witness, not a performance SLA").
- Iteration 1 is the staging note with its banner. Mining-driven leaves refine one arc each. The final leaf removes the banner, regenerates the scope numbers at the cut head and runs the numbers-verify sweep last.
- Cross-family review before every iteration merges (§5).
Out of scope
- The cut, the version bump and the real-cut validation: #19652.
- What 13.3 contains: milestone v13.3 and the ROADMAP decide that.
- Brain and Institution notes, which belong to their own release lines.
- The 13.2 notes epic #14800 and its residuals.
- The blog stream: #13383.
Release Type and Stability: the operator's call (§4).
Avoided traps
- Writing the plan instead of the release. A cornerstone that misses the window moves to 13.4 (ROADMAP, "How it runs"), so the note describes what merged by the cut head. Cornerstone 2's last leaf (#19671, in review as #19711) is the live case.
- A lean changelog because 13.3 is a minor. Minors carry the same bar (§5).
- "A grid release." The theme is the grid, but the window also carries the first cut through a release PR and the first phase of the 13.2 launch. A headline arc is not the release's extent (§1, rule 4).
- Counting from the engine's content mirror. It froze in August. §2's corpus run counts the window.
- A leaf that bundles drafting with cut mechanics.
Related
#19642 (ROADMAP 13.3) · #19652 (the cut) · #14800 (the 13.2 notes) · #14790 (the launch playbook) · milestone v13.3
Steward: @neo-opus-ada. The content-sized rows deposit is the first comment; reviews and the closing epic-resolution stay with the steward. Iteration leaves are claimable by any maintainer.
This epic absorbs #19714, which @neo-opus-vega filed 48 seconds later from the same operator prompt. The first filing survives per ticket-create §1a (ii), and #19714's unique substance is folded in above: the pre-flight message, the cross-line counts, the pictures bar, the candidate list, the bounded-witness bound, and the out-of-scope and avoided-traps additions.
Live latest-open sweep: latest 20 open issues at 2026-10-13T14:02Z; no equivalent found.
A2A sweep: latest 30 messages at 2026-10-13T14:02Z; no claim on the 13.3 notes.
MC sweep: "13.3 release notes missing, publish.mjs pre-flight errors when the staging note is absent", 4 results, no prior decision found.
Own-assignment sweep: 6 open, none overlapping (#14790 is the launch beside the gate).
Epic sweep: 25 open epics read, no shared terminal predicate (#14800 finishes 13.2's notes).
Structure map: N/A (a release-notes epic; no Agent OS placement claim).
Retrieval Hint: "content-sized rows", "header menus value picker", "publish.mjs release PR", "13.3 release notes", "The Grid Fits Its Content"
Terminal predicate:
.github/RELEASE_NOTES/v13.3.0.mdtells what 13.3 shipped to the release-notes bar and passes the §7 cut-readiness checklist at the cut head, so the #19652 cut publishes it.Problem
13.3 has no release notes and no epic for them.
publish.mjs preparestops on.github/RELEASE_NOTES/v13.3.0.md is missing(buildScripts/release/releaseState.mjs), and the cut record #19652 leaves the notes to this epic (release-notes workflow §6). The cut cannot start until the note exists.The window is short and fast. 13.2.0 was published at 2026-10-11T22:37Z, the ROADMAP (#19642) sets a one-to-two-week cadence, and milestone v13.3 is due 2026-10-25. By 2026-10-13T14:02Z, GitHub search counted 34 merged PRs in neo, 11 in neo-stage, 6 in Brain and 0 in Institution. These are search counts; the note's scope comes from the corpus run at the cut (§2). The arcs are hours old, so mining them costs less now than it ever will again (#14800's rationale).
The bar includes pictures. The operator's verdict on 13.2's iteration-6 draft was "not a single screenshot, not compelling" (#14800), and the published 13.2.0 note carries nine images.
Why an epic: the workflow (§1) makes release notes an epic by design. Iteration 1 merges early behind a banner, mining-driven leaves refine one arc each, and the last leaf passes §7. Each is a one-PR leaf with its own cross-family review.
Intended solution
dev-push burn-in), the Portal's featured Workstation reaching the package (#19640), and the launch running beside the gate (#14790).Out of scope
Release TypeandStability: the operator's call (§4).Avoided traps
Related
#19642 (ROADMAP 13.3) · #19652 (the cut) · #14800 (the 13.2 notes) · #14790 (the launch playbook) · milestone v13.3
Steward: @neo-opus-ada. The content-sized rows deposit is the first comment; reviews and the closing
epic-resolutionstay with the steward. Iteration leaves are claimable by any maintainer.This epic absorbs #19714, which @neo-opus-vega filed 48 seconds later from the same operator prompt. The first filing survives per
ticket-create§1a (ii), and #19714's unique substance is folded in above: the pre-flight message, the cross-line counts, the pictures bar, the candidate list, the bounded-witness bound, and the out-of-scope and avoided-traps additions.Live latest-open sweep: latest 20 open issues at 2026-10-13T14:02Z; no equivalent found. A2A sweep: latest 30 messages at 2026-10-13T14:02Z; no claim on the 13.3 notes. MC sweep: "13.3 release notes missing, publish.mjs pre-flight errors when the staging note is absent", 4 results, no prior decision found. Own-assignment sweep: 6 open, none overlapping (#14790 is the launch beside the gate). Epic sweep: 25 open epics read, no shared terminal predicate (#14800 finishes 13.2's notes). Structure map: N/A (a release-notes epic; no Agent OS placement claim).
Retrieval Hint: "content-sized rows", "header menus value picker", "publish.mjs release PR", "13.3 release notes", "The Grid Fits Its Content"