Plan post-v18 release handoff and next goalpost

docs/BEARING.md

@@ -39,19 +39,23 @@ of handwritten adapter folklore.
 
 Current branch state at this boundary:
 
-- Current branch: `v18-release-prep-slices-97-102`
+- Current branch: `v18-release-prep-slices-103-112`
 - Base branch: `main`
-- Current `origin/main`: `023c7d75`
-- Latest merged PR: #106, v18 release-candidate evidence and review cleanup
-- Latest released package line: `17.0.1`
-- Release-prep package metadata: `18.0.0`, not tagged or published yet
+- Current `origin/main`: `fdcb5da9`
+- Latest merged PR: #107, v18 release-prep metadata and final gates
+- Latest released package line: `17.0.1` until the `v18.0.0` tag and publish
+  complete
+- Release-prep package metadata: `18.0.0`, merged to `main`, not tagged or
+  published yet
 - Latest completed implementation cycle:
-  `0250-v18-final-release-prep-replan`
-- Current work: release-prep PR review for `v18.0.0`; local release preflight
-  is green for `18.0.0` metadata, and the remaining public-release gates are
-  PR review, GitHub CI, merge to `main`, tag, and publish.
-- Cleanup checkpoint: PR #106 is merged to `main`; this branch starts from that
-  merge and must not widen the v18 promise while preparing the public tag.
+  `0260-next-goalpost-replan-after-v18-merge`
+- Current work: post-merge release handoff and next-goalpost planning. PR
+  review, GitHub CI, and merge are complete for the release-prep branch. The
+  remaining public-release operation is to tag and publish `v18.0.0` from
+  aligned `main`.
+- Cleanup checkpoint: PR #107 is merged to `main`; this branch starts from
+  that merge and must not widen the v18 promise while setting the next
+  engineering goalpost.
 
 The current v18 graph-model posture is:
 
@@ -155,15 +159,16 @@ The current v18 graph-model posture is:
   command-owned readings, emits the command report, and permits live-ref
   finalization only through a reviewed JSON request that matches observed
   runtime evidence.
-- V18 release-candidate blockers are now explicit: GitHub CI, PR review,
-  post-merge tag/publish work, and the accepted residual raw content/property
-  risk.
+- V18 release-candidate blockers are now narrowed: PR review, GitHub CI, and
+  merge are complete; post-merge tag/publish work remains, with accepted
+  residual raw content/property risk and the full-graph-streaming non-claim
+  preserved.
 
 That is useful progress, not a finish line. The repo now has migration safety,
 wet-run proof, guarded finalization, generated Continuum contract tie-back,
-release-candidate docs, `18.0.0` package metadata, and a green local release
-preflight. Public v18 still needs PR review, GitHub CI, merge to `main`, tag,
-and publish.
+release-candidate docs, `18.0.0` package metadata on `main`, a green local
+release preflight from the release-prep branch, and green GitHub CI on PR
+#107. Public v18 still needs tag and publish work from aligned `main`.
 
 ## What Just Shipped
 
@@ -230,6 +235,16 @@ PR #106 landed v18 slices 66 through 96:
 - review follow-up that tightened runtime replay validation, finalization JSON
   evidence binding, fixture coverage, and migration script structure.
 
+PR #107 landed v18 slices 97 through 102:
+
+- release-prep baseline and gate evidence;
+- residual raw content/property storage risk decision;
+- public operator release notes and v18 non-goals;
+- `18.0.0` package, JSR, lockfile, workspace, changelog, and release-policy
+  metadata alignment;
+- final release-prep replan before PR review;
+- technical teardown publication and review follow-up.
+
 Slice 36 is complete on this branch. The migration manifest root now exists
 as a frozen domain noun, with runtime-backed basis, mapping, warning, and
 fatal-error entries. It does not serialize, read Git, or write graph history.
@@ -387,13 +402,14 @@ and wet-run fixture harnessing.
   map because historical replay tests carry pre-codec inline fixture classes
   that are not `PropValue`-honest enough for `LegacyPropertyValue`.
 - The v18 migration tool now replays migrated scratch operations through the
-  production graph runtime write/materialization path during wet runs, but
-  public release still needs the full release-prep gate set on the eventual
-  release branch.
+  production graph runtime write/materialization path during wet runs, and the
+  release-prep PR has merged. Public release now needs tag and publish
+  evidence from aligned `main`.
 - Legacy readings from the v17 golden fixture now have restored public-read
   construction, but broader non-fixture replay coverage remains future work.
 - The command wrapper can finalize through a reviewed JSON request, but the
-  public tag still needs CI and release-preflight validation.
+  public tag still needs aligned-main release preflight, tag evidence, and
+  registry publish evidence.
 - Continuum/WARP Optic release claims now have generated contract evidence for
   runtime-boundary graph-model migration, but native Continuum witnesshood is
   still not claimed.
@@ -405,21 +421,17 @@ and wet-run fixture harnessing.
 
 ## Where We Are Heading
 
-The remaining runway is no longer a five-slice tail. The next realistic plan
-is thirty slices. Some slices may collapse when evidence is in hand, but the
-release plan should assume the proof work is hard until it is proven easy.
+The release-prep branch has merged. The next move is no longer adding v18
+features. It is one of three explicitly separated actions:
 
-The first thirty slices converted operation-derived confidence into
-production-runtime confidence, restored the v17 golden fixture, drove the
-canonical wet-run report to zero public-read mismatches, guarded live
-finalization, tied graph-model migration to generated runtime-boundary
-contracts, and added the first `warp-ttd` generated-family smoke. The next
-goalpost is release-prep hardening rather than new feature expansion.
+1. cut and publish `v18.0.0` from aligned `main`;
+2. retire one more raw content/property compatibility boundary;
+3. start the v19 observer/admission witnesshood runway.
 
-Slice 96 reconciled the backlog ledger with that evidence. The next practical
-goalpost is a release-prep branch that runs the full gate set, freezes package
-and release notes, and decides whether remaining raw content/property storage
-retirement blocks the public tag or ships as explicit residual risk.
+Those should not be blended. Tag/publish is release operation. Storage
+retirement is substrate-debt reduction. V19 starts a new semantic runway.
+
+Slices 103 through 112 record that split and keep the next goalpost honest.
 
 ### Release-Prep Checklist
 
@@ -436,6 +448,29 @@ retirement blocks the public tag or ships as explicit residual risk.
 - [x] 102. Replan from final release-prep evidence before PR review:
   [0250](design/0250-v18-final-release-prep-replan/v18-final-release-prep-replan.md).
 
+### Post-Merge Goalpost Checklist
+
+- [x] 103. Record the v18 post-merge release handoff:
+  [0251](design/0251-v18-post-merge-release-handoff/v18-post-merge-release-handoff.md).
+- [x] 104. Define the v18 tag and publish gate:
+  [0252](design/0252-v18-tag-publish-gate/v18-tag-publish-gate.md).
+- [x] 105. Define the v18 release evidence archive:
+  [0253](design/0253-v18-release-evidence-archive/v18-release-evidence-archive.md).
+- [x] 106. Frame the post-v18 storage retirement decision:
+  [0254](design/0254-post-v18-storage-retirement-decision/post-v18-storage-retirement-decision.md).
+- [x] 107. Frame the v19 native Continuum witnesshood runway:
+  [0255](design/0255-v19-native-continuum-witnesshood-runway/v19-native-continuum-witnesshood-runway.md).
+- [x] 108. Scope end-to-end graph streaming reads and writes to v20:
+  [0256](design/0256-v20-streaming-reads-writes-scope/v20-streaming-reads-writes-scope.md).
+- [x] 109. Define the post-v18 public doc honesty audit:
+  [0257](design/0257-post-v18-public-doc-honesty-audit/post-v18-public-doc-honesty-audit.md).
+- [x] 110. Define the next residual boundary retirement decision:
+  [0258](design/0258-v18-residual-boundary-next-retirement/v18-residual-boundary-next-retirement.md).
+- [x] 111. Plan backlog lane cleanup after v18:
+  [0259](design/0259-backlog-lane-cleanup-after-v18/backlog-lane-cleanup-after-v18.md).
+- [x] 112. Replan the next goalpost after the v18 release-prep merge:
+  [0260](design/0260-next-goalpost-replan-after-v18-merge/next-goalpost-replan-after-v18-merge.md).
+
 ### Next Thirty-Slice Checklist
 
 - [x] 66. Design production-runtime scratch replay conformance:
@@ -799,3 +834,23 @@ and concrete checks live in `docs/invariants/`.
   [0249](design/0249-v18-version-tag-readiness/v18-version-tag-readiness.md).
 - [x] 102. Replan from final release-prep evidence before PR review:
   [0250](design/0250-v18-final-release-prep-replan/v18-final-release-prep-replan.md).
+- [x] 103. Record the v18 post-merge release handoff:
+  [0251](design/0251-v18-post-merge-release-handoff/v18-post-merge-release-handoff.md).
+- [x] 104. Define the v18 tag and publish gate:
+  [0252](design/0252-v18-tag-publish-gate/v18-tag-publish-gate.md).
+- [x] 105. Define the v18 release evidence archive:
+  [0253](design/0253-v18-release-evidence-archive/v18-release-evidence-archive.md).
+- [x] 106. Frame the post-v18 storage retirement decision:
+  [0254](design/0254-post-v18-storage-retirement-decision/post-v18-storage-retirement-decision.md).
+- [x] 107. Frame the v19 native Continuum witnesshood runway:
+  [0255](design/0255-v19-native-continuum-witnesshood-runway/v19-native-continuum-witnesshood-runway.md).
+- [x] 108. Scope end-to-end graph streaming reads and writes to v20:
+  [0256](design/0256-v20-streaming-reads-writes-scope/v20-streaming-reads-writes-scope.md).
+- [x] 109. Define the post-v18 public doc honesty audit:
+  [0257](design/0257-post-v18-public-doc-honesty-audit/post-v18-public-doc-honesty-audit.md).
+- [x] 110. Define the next residual boundary retirement decision:
+  [0258](design/0258-v18-residual-boundary-next-retirement/v18-residual-boundary-next-retirement.md).
+- [x] 111. Plan backlog lane cleanup after v18:
+  [0259](design/0259-backlog-lane-cleanup-after-v18/backlog-lane-cleanup-after-v18.md).
+- [x] 112. Replan the next goalpost after the v18 release-prep merge:
+  [0260](design/0260-next-goalpost-replan-after-v18-merge/next-goalpost-replan-after-v18-merge.md).

docs/design/0251-v18-post-merge-release-handoff/v18-post-merge-release-handoff.md

@@ -0,0 +1,46 @@
+# V18 Post-Merge Release Handoff
+
+## Hill
+
+Record the state after the release-prep PR merged and make the remaining
+`v18.0.0` release work explicit: tag and publish from `main`, with no new
+runtime promise added on a follow-up branch.
+
+## Context
+
+PR #107 merged the `18.0.0` metadata, release notes, release-prep evidence,
+and technical teardown to `main`. The package line is release-ready in source,
+but the public release is not complete until the tag and publish artifacts are
+cut from merged `main`.
+
+The follow-up branch exists to cleanly bookkeep that handoff and set the next
+engineering goalpost. It must not blur three different states:
+
+- `18.0.0` metadata merged to `main`;
+- `v18.0.0` tag created from `main`;
+- npm and JSR artifacts published from that release path.
+
+## User Stories
+
+- As a release manager, I can tell whether the source tree, the Git tag, and
+  the package registries all agree on `18.0.0`.
+- As a maintainer, I can see that the release-prep PR has merged without
+  accidentally treating that as a completed public release.
+- As a downstream operator, I can inspect release notes that describe actual
+  `v18` guarantees without inheriting hidden branch-state assumptions.
+
+## Acceptance Criteria
+
+- `docs/BEARING.md` names PR #107 as merged and `origin/main` as the source
+  of truth for the current branch.
+- The v18 blocker ledger treats PR review, CI, and merge as complete.
+- Tag and publish remain explicitly open until registry evidence exists.
+- No follow-up doc claims that end-to-end graph streaming is part of v18.
+
+## Test Plan
+
+- `git status --short --branch`
+- `git rev-parse --short origin/main`
+- `gh pr view 107 --json state,mergedAt,mergeCommit`
+- `npx markdownlint` on edited Markdown files
+- `git diff --check`

docs/design/0252-v18-tag-publish-gate/v18-tag-publish-gate.md

@@ -0,0 +1,43 @@
+# V18 Tag And Publish Gate
+
+## Hill
+
+Define the final public-release gate for `v18.0.0` without performing it on a
+feature branch.
+
+## Context
+
+`package.json`, `jsr.json`, workspace package manifests, lockfile metadata,
+and the changelog now agree on `18.0.0`. The release-prep branch passed local
+preflight before merge, and GitHub CI passed before PR #107 merged.
+
+That is necessary but not sufficient. The public release must still be cut
+from merged `main`, and publish evidence must agree with the tag target.
+
+## User Stories
+
+- As a release manager, I can run a short gate before tagging and know which
+  evidence must be attached to the release record.
+- As a package consumer, I can trust that the npm artifact, JSR artifact, and
+  Git tag describe the same source tree.
+- As a maintainer, I do not have to infer whether a feature branch is allowed
+  to publish.
+
+## Acceptance Criteria
+
+- The release gate requires `main` to match `origin/main`.
+- The gate requires `npm run release:preflight` on `main` before tagging.
+- The gate requires the tag target, npm dry-run, JSR dry-run, and published
+  artifacts to agree on `18.0.0`.
+- The gate does not create a tag from a feature branch.
+- The gate preserves the v18 non-claim around full graph streaming reads and
+  writes.
+
+## Test Plan
+
+- `git checkout main`
+- `git pull --ff-only origin main`
+- `npm run release:preflight`
+- `git tag --list v18.0.0`
+- `npm view @git-stunts/git-warp version`
+- JSR package version inspection after publish

docs/design/0253-v18-release-evidence-archive/v18-release-evidence-archive.md

@@ -0,0 +1,49 @@
+# V18 Release Evidence Archive
+
+## Hill
+
+Name the evidence that should survive after the public `v18.0.0` release,
+instead of leaving it scattered across PR comments, CI logs, and local command
+output.
+
+## Context
+
+The v18 path accumulated several classes of release evidence:
+
+- migration wet-run reports;
+- generated Continuum contract conformance;
+- `warp-ttd` generated-family smoke evidence;
+- closeout audit evidence for raw content/property compatibility boundaries;
+- release preflight output;
+- GitHub CI checks;
+- npm and JSR publish output.
+
+PR comments and CI logs are useful, but they are not a durable release ledger
+by themselves. A release evidence archive should summarize the final public
+facts and point back to inspectable sources.
+
+## User Stories
+
+- As a maintainer, I can answer "what exactly shipped in v18?" without
+  replaying the whole PR history.
+- As a downstream integrator, I can find the migration proof and generated
+  contract evidence that justify the release claim.
+- As a future release agent, I can compare v19 or v20 release evidence against
+  a stable v18 pattern.
+
+## Acceptance Criteria
+
+- A future release archive names the tag, commit, package versions, publish
+  timestamps, and registry artifacts.
+- The archive links to the release-prep PR, CI run, release notes, and local
+  gate evidence.
+- The archive records residual raw content/property compatibility risk as an
+  accepted v18 risk, not an omitted blocker.
+- The archive records that full graph streaming is a later-major goal.
+
+## Test Plan
+
+- Markdown lint the archive.
+- Link-check repository-local release evidence links.
+- Compare archived package versions against `package.json` and `jsr.json`.
+- Verify the archived tag SHA matches the release commit.

docs/design/0254-post-v18-storage-retirement-decision/post-v18-storage-retirement-decision.md

@@ -0,0 +1,43 @@
+# Post-V18 Storage Retirement Decision
+
+## Hill
+
+Decide whether the next engineering goalpost should continue raw
+content/property storage retirement before starting v19 runtime doctrine work.
+
+## Context
+
+V18 deliberately accepts residual raw content/property compatibility
+boundaries. That is honest, but it is still debt. The closeout audit now
+guards drift, yet the repo still carries legacy compatibility spelling in
+named boundaries.
+
+The competing pressure is v19: observation, admission, and doctrine
+convergence. Starting v19 while storage-plane debt remains visible may be
+reasonable, but only if the debt is boxed and ratcheted tightly enough.
+
+## User Stories
+
+- As a maintainer, I can decide whether storage debt is a release-line risk or
+  a manageable ratcheted backlog item.
+- As a migration operator, I can trust that remaining legacy compatibility
+  boundaries are named and tested.
+- As a v19 planner, I can start observer/admission work without inheriting
+  ambiguous v18 substrate claims.
+
+## Acceptance Criteria
+
+- The decision distinguishes total storage retirement from bounded residual
+  compatibility.
+- The decision identifies the next single raw boundary to retire if retirement
+  continues first.
+- The decision explains why v19 can or cannot start before full storage
+  retirement.
+- The closeout audit remains the executable guard for any deferred boundary.
+
+## Test Plan
+
+- Run the content/property closeout audit.
+- Inspect each allowlisted boundary for ownership and reason.
+- Verify release docs do not claim total raw storage retirement.
+- Update the relevant backlog lane after the decision.