diff --git a/DESIGN/01-storage-format.md b/DESIGN/01-storage-format.md
index e5fb758..3de6177 100644
--- a/DESIGN/01-storage-format.md
+++ b/DESIGN/01-storage-format.md
@@ -91,7 +91,7 @@ Schema-owned display fields are Lanework's to interpret — **coerce where a sen
 
 Future integrations extend the core schema; this version reserves their names — and, for comments, specifies the storage schema ahead of the feature — so nothing else squats on them. The app treats reserved keys as ordinary unknown keys (preserved verbatim, invisible in the UI) and reserved folders as ignored non-level content:
 
-- **`comments/` card child — the comment thread** (storage specified 2026-07-29, the card-window UI designed the same day — 05-card-window.md ▸ The comments column; no comments functionality ships in this version). A comment is a **UUID-named folder** under `comments/`, holding `index.md` (frontmatter + Markdown body, the card-body subset) and optionally `attachments/` — a card's anatomy one level down, so the fractal rules apply verbatim: two-step-create tolerance, stray tolerance, loose-file relocation, displacement of a squatted `attachments`, shape-only identity. **The heals run window-scoped at this depth** (ruled 2026-07-31): comments are outside the board walk, so the thread read is the detector — it reports loose files and claimed-name squatters as window-scoped defects, and the card window relocates/displaces them memo-less on the read-cadence guard (02-architecture.md ▸ HealScheduler's window-scoped pattern); a loose file beside a comment's `index.md` moves into that comment's `attachments/` with the relocation family's notice ("Moved 'notes.txt' into attachments — comment on 'Fix login'"), and the heals compose as at board level. `comments/` itself stays a plain reserved child — never a level, never identity. **Field table for `kind: comment`** (exercising IntegrityRules' per-kind hook for the first time): `schema` required; `kind: comment` written at creation (the common rule); **`author`** lenient — self-reported *content*, not overlay: unlike `modified-by` it survives app writes; the app writes the macOS account's full name (the identity 06-history-undo.md's derived default already uses), agents write their own (the guide teaches it), tracker sync writes the remote author verbatim, and missing renders unattributed; `created`/`modified` as everywhere — `created` is load-bearing (ordering below), and the edited indicator is simply `modified` differing, no extra field; `remote`/`remote-state` ride as reserved tracker keys. **No `title`, no `order`.** The rules, all ruled 2026-07-29: **Ordering is chronology, not ranks** — the thread sorts by `created` ascending; chronology *is* a conversation's semantics, and cross-machine or tracker-synced comments carry independent clocks where minted ranks would interleave arbitrarily (the trash joined the timestamp family 2026-07-31 — it sorts by `modified` descending, Deletion below); ties and missing/malformed `created` (coerce-tier fallback, logged) sort after dated siblings, folder-name order. **Flat this iteration** — no reply nesting, deliberately deferred with the door open (a future reply relation would be frontmatter, and the window-scoped read below holds either way). **Comment defects never refuse the board** — worst case is the stray posture (tolerated, logged, unrendered): a broken leaf annotation must not brick a load; deliberate, proportionate divergence from card fail-fast. The posture is **load-scoped** (ruled 2026-07-31): a user-initiated copy is a transaction whose stamping preflight refuses whole at comment depth too (Identity lifecycle above) — the board always loads; the gesture may refuse. **Delete skips the board trash, not the pattern** (re-ruled 2026-07-29 — supersedes the earlier capture parenthetical): the root `.trash/` holds cards and lanes; a deleted comment **moves into `comments/.trash/`** — the materialized-trash pattern one level down, joining `.draft` in the claimed names (a squatter displaces by the ladder). The move is immediate, no confirm, and undo is the net: the inverse is the ordinary move back — **no byte capture in any tier** (13-native-undo.md's no-capture rule holds; on git boards the move is a commit like any other). `comments/.trash/` is undo's backing store, never a UI surface: excluded from the thread, **purged when undo no longer needs it** (re-ruled 2026-07-31 — the session-coarsening model, 13-native-undo.md: on free-tier boards the close step's undo restores deleted comments, so the purge waits until that coarse step leaves the board stack or the board session ends; on Pro the close commit nets delete-plus-purge and revert restores from history, so the purge rides the close flush; crash residue sweeps at the next card-window open, armed-then-cleared like every heal memo — residue defined as content no live coarse step owns, ruled 2026-07-31: 13-native-undo.md's sweep gate), and **stripped at every copy boundary** (clipboard staging, Duplicate, Save as Template — a copy must not carry ghosts no window session will ever purge; sibling of the `remote`-severing rule). The container-change stamping rule applies — the move stamps `modified` (`modified`'s scope above). **Copies carry the thread** — a copy is a fork, and dropping a subtree would be the one place a copy loses content; comment folders remint like every copied folder, and the copy's comments drop `remote`/`remote-state` (an instance of the item-level sever rule — Identity lifecycle above); whole-board forks carry verbatim as always. **Comments are window-scoped, outside the board snapshot** — the walk stays O(cards): the card window reads its own thread and live-reloads it from the same FSEvents stream, and the board snapshot never loads comment content. This is the stated exception to snapshot completeness, costs named: the snapshot never carries comment content — board search reaches comment bodies through its own transient search-time sweep instead (04-interactions.md ▸ Search, re-ruled 2026-07-29), the card window's find covers the focused thread (05) — and foreign comment changes are described by **path shape** — the "Update agent guide (vN)" mechanism: a changed path under `…/comments/<uuid>/` composes "Comment on '⟨card title⟩'" / "Edit comment on…" / "Delete comment on…" — and a comment leaving the tree from `comments/.trash/` composes **"Permanently delete comment on…"** (blessed 2026-07-31 — the board trash-pair's leaving-the-tree rule one level down, so purge-only windows never fall to a shrug and foreign purges read honestly); plurals fold noun-first where the post subject forces it (blessed 2026-07-31): "3 comments on 'Fix login'" / "Comment on 3 cards" / "Edit 3 comments on 'Fix login'" / "Draft comment on 3 cards" — (06 gains the verb family when the feature ships), and the announcer speaks arrivals the same way. Git protection needs nothing new — the committer already stages the whole root and its condition is the tree. Trash interplay is free: a trashed card carries its `comments/`, purge kills it, restore returns it. Comments never count in the welcome counts. A comment-count chip on card faces is WISHLIST #9. **`comments/.draft/` is the card's single comment draft** (ruled 2026-07-29, the UI pass — user content with a lifecycle name, the `.trash` pattern one level down): a reserved dot-named folder holding ordinary comment schema, `attachments/` included; excluded from the thread; **posting renames it to a fresh lowercase UUID and restamps `created`/`modified` in the same bracket** — chronology is post time, not drafting time — one commit. `.draft` joins the claimed names (a wrong-kind node squatting it displaces by the ladder); copies and the trash carry it like any comment folder (fork-lossless); a draft emptied of text with no attachments is deleted by the app, never litter. The schema is **tier-blind** (one format, no tier-specific keys), and so is the feature: **comments render and author in every tier** (ruled 2026-07-29 — 05-card-window.md ▸ The comments column); only tracker sync is tier-gated (Teams — 12-editions.md).
+- **`comments/` card child — the comment thread** (storage specified 2026-07-29, the card-window UI designed the same day — 05-card-window.md ▸ The comments column; no comments functionality ships in this version). A comment is a **UUID-named folder** under `comments/`, holding `index.md` (frontmatter + Markdown body, the card-body subset) and optionally `attachments/` — a card's anatomy one level down, so the fractal rules apply verbatim: two-step-create tolerance, stray tolerance, loose-file relocation, displacement of a squatted `attachments`, shape-only identity. **The heals run window-scoped at this depth** (ruled 2026-07-31): comments are outside the board walk, so the thread read is the detector — it reports loose files and claimed-name squatters as window-scoped defects, and the card window relocates/displaces them memo-less on the read-cadence guard (02-architecture.md ▸ HealScheduler's window-scoped pattern); a loose file beside a comment's `index.md` moves into that comment's `attachments/` with the relocation family's notice ("Moved 'notes.txt' into attachments — comment on 'Fix login'"), and the heals compose as at board level. `comments/` itself stays a plain reserved child — never a level, never identity. **Field table for `kind: comment`** (exercising IntegrityRules' per-kind hook for the first time): `schema` required; `kind: comment` written at creation (the common rule); **`author`** lenient — self-reported *content*, not overlay: unlike `modified-by` it survives app writes; the app writes the macOS account's full name (the identity 06-history-undo.md's derived default already uses), agents write their own (the guide teaches it), tracker sync writes the remote author verbatim, and missing renders unattributed; `created`/`modified` as everywhere — `created` is load-bearing (ordering below), and the edited indicator is simply `modified` differing, no extra field; `remote`/`remote-state` ride as reserved tracker keys. **No `title`, no `order`.** The rules, all ruled 2026-07-29: **Ordering is chronology, not ranks** — the thread sorts by `created` ascending; chronology *is* a conversation's semantics, and cross-machine or tracker-synced comments carry independent clocks where minted ranks would interleave arbitrarily (the trash joined the timestamp family 2026-07-31 — it sorts by `modified` descending, Deletion below); ties and missing/malformed `created` (coerce-tier fallback, logged) sort after dated siblings, folder-name order. **Flat this iteration** — no reply nesting, deliberately deferred with the door open (a future reply relation would be frontmatter, and the window-scoped read below holds either way). **Comment defects never refuse the board** — worst case is the stray posture (tolerated, logged, unrendered): a broken leaf annotation must not brick a load; deliberate, proportionate divergence from card fail-fast. The posture is **load-scoped** (ruled 2026-07-31): a user-initiated copy is a transaction whose stamping preflight refuses whole at comment depth too (Identity lifecycle above) — the board always loads; the gesture may refuse. **Delete skips the board trash, not the pattern** (re-ruled 2026-07-29 — supersedes the earlier capture parenthetical): the root `.trash/` holds cards and lanes; a deleted comment **moves into `comments/.trash/`** — the materialized-trash pattern one level down, joining `.draft` in the claimed names (a squatter displaces by the ladder). The move is immediate, no confirm, and undo is the net: the inverse is the ordinary move back — **no byte capture in any tier** (13-native-undo.md's no-capture rule holds; on git boards the move is a commit like any other). `comments/.trash/` is undo's backing store, never a UI surface: excluded from the thread, **purged when undo no longer needs it** (re-ruled 2026-07-31 — the session-coarsening model, 13-native-undo.md: on free-tier boards the close step's undo restores deleted comments, so the purge waits until that coarse step leaves the board stack cleanly or the board session ends (a stale-skipped step's backing waits for session end — 13's skip-purge decoupling, ruled 2026-07-31); on Pro the close commit nets delete-plus-purge and revert restores from history, so the purge rides the close flush; crash residue sweeps at the next card-window open, armed-then-cleared like every heal memo — residue defined as content no live coarse step owns, ruled 2026-07-31: 13-native-undo.md's sweep gate), and **stripped at every copy boundary** (clipboard staging, Duplicate, Save as Template — a copy must not carry ghosts no window session will ever purge; sibling of the `remote`-severing rule). The container-change stamping rule applies — the move stamps `modified` (`modified`'s scope above). **Copies carry the thread** — a copy is a fork, and dropping a subtree would be the one place a copy loses content; comment folders remint like every copied folder, and the copy's comments drop `remote`/`remote-state` (an instance of the item-level sever rule — Identity lifecycle above); whole-board forks carry verbatim as always. **Comments are window-scoped, outside the board snapshot** — the walk stays O(cards): the card window reads its own thread and live-reloads it from the same FSEvents stream, and the board snapshot never loads comment content. This is the stated exception to snapshot completeness, costs named: the snapshot never carries comment content — board search reaches comment bodies through its own transient search-time sweep instead (04-interactions.md ▸ Search, re-ruled 2026-07-29), the card window's find covers the focused thread (05) — and foreign comment changes are described by **path shape** — the "Update agent guide (vN)" mechanism: a changed path under `…/comments/<uuid>/` composes "Comment on '⟨card title⟩'" / "Edit comment on…" / "Delete comment on…" — and a comment leaving the tree from `comments/.trash/` composes **"Permanently delete comment on…"** (blessed 2026-07-31 — the board trash-pair's leaving-the-tree rule one level down, so purge-only windows never fall to a shrug and foreign purges read honestly); plurals fold noun-first where the post subject forces it (blessed 2026-07-31): "3 comments on 'Fix login'" / "Comment on 3 cards" / "Edit 3 comments on 'Fix login'" / "Draft comment on 3 cards" — (06 gains the verb family when the feature ships), and the announcer speaks arrivals the same way. Git protection needs nothing new — the committer already stages the whole root and its condition is the tree. Trash interplay is free: a trashed card carries its `comments/`, purge kills it, restore returns it. Comments never count in the welcome counts. A comment-count chip on card faces is WISHLIST #9. **`comments/.draft/` is the card's single comment draft** (ruled 2026-07-29, the UI pass — user content with a lifecycle name, the `.trash` pattern one level down): a reserved dot-named folder holding ordinary comment schema, `attachments/` included; excluded from the thread; **posting renames it to a fresh lowercase UUID and restamps `created`/`modified` in the same bracket** — chronology is post time, not drafting time — one commit. `.draft` joins the claimed names (a wrong-kind node squatting it displaces by the ladder); copies and the trash carry it like any comment folder (fork-lossless); a draft emptied of text with no attachments is deleted by the app, never litter. The schema is **tier-blind** (one format, no tier-specific keys), and so is the feature: **comments render and author in every tier** (ruled 2026-07-29 — 05-card-window.md ▸ The comments column); only tracker sync is tier-gated (Teams — 12-editions.md).
 - **Card metadata keys `labels`, `assignees`, `due`** — issue-style metadata belonging to the tracker-integration story (the pathfinder shipped them natively; the rewrite reserves them instead). One deliberate carve-out: the commit-message composer names changes to these three keys ("Relabel card 'X'" — 06-history-undo.md), invisible in the UI though they are.
 - **`remote` key** (board, card) and **`remote-state` key** (lane) — future tracker connectors (Gitea/GitHub/GitLab/Jira), per 07-sync-collab.md's out-of-scope note.
 
diff --git a/DESIGN/06-history-undo.md b/DESIGN/06-history-undo.md
index c4e1715..3f6173e 100644
--- a/DESIGN/06-history-undo.md
+++ b/DESIGN/06-history-undo.md
@@ -12,7 +12,7 @@ Git is the undo substrate — on boards that have git. **Git is opt-in per board
 - **A `.git` that isn't a valid repository still reads as git mode — and fails loudly** (ruled 2026-07-31): detection is presence-shaped (any root `.git` entry, directory or worktree/submodule pointer file), so a corrupt or unopenable repo never falls to mode none — Add Git is never offered against an existing `.git`, whatever its condition (init into a repairable repo is exactly the never-mutate hazard). The board itself loads and edits normally — files are the board — but the failure is **loud**: a standing breakage-class banner at detection ("This board's git repository can't be read — history is paused; Lanework leaves the repository untouched"), announced per 10-accessibility.md, with the whole git surface paused (the abnormal-states posture below) and the popover's git section naming the state; the banner clears when a later open or reload finds the repo readable. Never a silent placeholder discovered only in the popover.
 - **Abnormal repo states** (settled; adoption never assumes a tidy clone): an **unborn HEAD** (`git init`, no commits yet) is normal git mode — the first auto-commit creates the root commit on the branch HEAD names, and the undo trail simply starts empty. **The root commit has its own subject** (settled): whenever the app creates a repo's first commit — immediately on the app's own add-git (init doesn't wait for the debounce; the board is protected from the moment git exists), or at the first settled change on an adopted unborn repo — it commits the whole tree as **"Initial board state"**, never a folded diff-from-empty: there is no last-committed snapshot to diff against, and forty Adds would bury the event. **The root commit is never split and is user-authored** (blessed 2026-07-31): it is a baseline, not a change-set — the event it records is the user's act of putting the board under git, adopted unborn repos' unwitnessed files included. A **detached HEAD**, or an **in-progress merge/rebase/cherry-pick** left by outside-the-app git (`MERGE_HEAD`, `rebase-merge`/`rebase-apply`, `CHERRY_PICK_HEAD` — pause states that load fine on a clean tree and are otherwise invisible), instead **pauses the git surface honestly — the *whole* surface, remote half included** (settled): auto-commit holds (the auth-pause posture, 07-sync-collab.md — pause, badge, explain, never hammer), Undo/Redo and the branch controls disable, **and Pull, Push, and push-on-commit hold with them** — with auto-commit held, 07's clean-tree-by-pull-time invariant is false, and a pull's rebase (or a rejected push's fetch→rebase→push) would run against a dirty tree carrying uncommitted edits, precisely what flush-before-overwrite exists to prevent; the ahead/behind badge keeps counting (a fetch is a read), and the popover's git section names the state plainly ("HEAD is detached — commits would belong to no branch"; "a merge is in progress") and says resolving it belongs to the tool that created it. Edits keep landing on disk — files are the board — and commit as one settled batch when the state clears. The app **never mutates repo state it didn't create** (no auto branch-at-HEAD, no `merge --abort`); the check runs at open and again before every flush — and, because a terminal's cleanup moves only files under `.git`, which the watcher never delivers, **a standing pause re-reads the repository state every 15 s** (blessed 2026-07-31; injectable cadence, only while paused, a handful of stats — not a retry, nothing is attempted) — so finishing the operation in a terminal resumes the pipeline without ceremony. **The one exemption is the app's own leftovers** (settled): every bracketed operation stamps its intent app-side (per-board registry) before touching the repo, so an interrupted app-run rebase or checkout is recognizable as Lanework's — finding a pause state with a matching stamp, the app **aborts its own unfinished operation** to restore the pre-operation state and says so via banner ("a branch switch was interrupted — the previous state is restored"), then clears the stamp — **on success only; a failed abort keeps the stamp** (blessed 2026-07-31): the stamp is the sole evidence the leftover is the app's, and clearing it on failure would demote the leftover to somebody-else's forever — the pause would then send the user to "the tool that created it," which was this app. Kept, the next open or paused-state re-read recognizes it and retries; recovery converges instead of orphaning. Abort discards nothing: fetched commits stay fetched, local commits are restored — the rebase's own no-loss accounting. Without a matching stamp the leftover is outside git's, and the pause-and-defer stance above holds unchanged.
 - **Boards nested inside an existing repository are left strictly alone** — git cannot be added to them (no nested repo, no commits into the user's repo), so they get **no app-managed history — while the native session stack still serves ⌘Z there in every tier** (re-ruled 2026-07-31, retiring the old no-undo residue: 13's stack is memory-only and journal-free, so self-containment holds; what repo-nested denies is git, never undo). The board popover's git section must say so honestly: not a hidden "add git" but a short explanation ("this board lives inside a repository; Lanework leaves it to that repository") — the option is absent because it *can't* apply, and the UI should teach that rather than look broken.
-- **Auto-commit**: every settled change (debounced past drag/typing churn) commits with a descriptive message ("Move card 'Fix login' to Doing"). Board history sees **card-window sessions, not gestures** (re-ruled 2026-07-31, widening the Edit-sessions-not-save-ticks rule — 13-native-undo.md's session-coarsening model, applied to the commit substrate): while a card's window is open, everything happening inside it — the body editor's ~700 ms crash-safe disk saves, comment posts and deletes, draft-save cadence, sidebar changes — stays **uncommitted**, and the committer **stages around the whole open card folder** (the former Edit-session stage-around, widened; comments included); **window close flushes the session as one commit** ("Update card 'Fix login'"-shaped, the composer folding the card-scoped diff, body bullets carrying the events) — granular window activity never litters history. The EchoLedger's two-commit split still applies at close when the held window mixes foreign changes to that card with the app's own. Settled-tree events that cannot wait — a pull's flush-before-overwrite — commit the session's on-disk state as-is (a mechanical exception; 07-sync-collab.md's same-card signpost covers the visible half); branch switch instead gates on explicit save-or-discard (Branch switching below). The cadence constraint below demands batching at least this coarse. **Board-window close and app quit flush the pipeline** — any pending editor save (05-card-window.md), then open card-window sessions, then the pending auto-commit — before teardown; nothing settled is ever left unsaved or uncommitted by closing.
+- **Auto-commit**: every settled change (debounced past drag/typing churn) commits with a descriptive message ("Move card 'Fix login' to Doing"). Board history sees **card-window sessions, not gestures** (re-ruled 2026-07-31, widening the Edit-sessions-not-save-ticks rule — 13-native-undo.md's session-coarsening model, applied to the commit substrate): while a card's window is open, everything happening inside it — the body editor's ~700 ms crash-safe disk saves, comment posts and deletes, draft-save cadence, sidebar changes — stays **uncommitted**, and the committer **stages around the whole open card folder** (the former Edit-session stage-around, widened; comments included); **window close flushes the session as one commit** ("Update card 'Fix login'"-shaped, the composer folding the card-scoped diff, body bullets carrying the events) — granular window activity never litters history. **The flush awaits the snapshot that covers it** (ruled 2026-07-31): the composer diffs `store.snapshot` against HEAD, so the close flush awaits a snapshot generation covering its changed paths before the committer runs — the commit's subject can never be outrun by its own reload; the cadence margin (2 s debounce vs 200 ms watcher) is the practical cushion, never the guarantee (the awaiting-bracket machinery is the shape). **A commit's comment bullets sort chronologically** (blessed 2026-07-31): by the comments' own `created`, folder name on ties — event order reads as the conversation did, never UUID-arbitrary. The EchoLedger's two-commit split still applies at close when the held window mixes foreign changes to that card with the app's own. Settled-tree events that cannot wait — a pull's flush-before-overwrite — commit the session's on-disk state as-is (a mechanical exception; 07-sync-collab.md's same-card signpost covers the visible half); branch switch instead gates on explicit save-or-discard (Branch switching below). The cadence constraint below demands batching at least this coarse. **Board-window close and app quit flush the pipeline** — any pending editor save (05-card-window.md), then open card-window sessions, then the pending auto-commit — before teardown; nothing settled is ever left unsaved or uncommitted by closing.
 - **Flush-before-overwrite**: before an app write overwrites on-disk state that differs from the last-loaded snapshot (an uncommitted external change — e.g. an agent's body rewrite racing the card editor's debounced save, 05-card-window.md), the pending auto-commit is flushed so the external version enters history first. "Both versions exist as commits" is thereby a guarantee, not a likelihood — **as strong as the watcher's knowledge** (bounded, blessed 2026-07-31): the gate fires on *known* foreign changes, learned from landed reloads, so a foreign write still inside the watcher debounce can be overwritten unflushed; the window is the debounce (~200 ms), the disk-level outcome inside it is 05's last-writer-wins — the same ruling at history granularity — and `index.md` writes compose over fresh disk bytes anyway, so wholesale loss needs a body-save or raw-Apply race. A **failed reload counts as a foreign change** (can't know, so protect). A pre-write disk compare was weighed and declined — a read per write to buy back the debounce corner. **The flush commits synchronously on the main actor** (blessed 2026-07-31): the write path it must precede is synchronous, so the ordering guarantee can only be kept by a synchronous commit — 02's hang-avoidance doctrine and this guarantee genuinely conflict here, and the guarantee wins, being rare (the known-foreign window only), bounded (one stage-and-commit plus the composer's HEAD-tree materialization — a few board-sized walks, accepted so exactly the commit that preserves somebody else's version keeps its full semantic subject), and load-bearing (the alternative loses a version no commit protects). One attempt, no lock backoff on this path — a held `index.lock` falls through to the next quiet debounce. The same flush settles the tree before a pull runs (07-sync-collab.md).
 - **Cadence constraint** (agreed): the auto-commit cadence must not make the history of a remote-shared board unbearable — one commit per drag is fine for a local undo trail but noisy as a shared log. The debounce/batching design here must serve both consumers; the push/pull side is settled in 07-sync-collab.md (optional push-on-commit, automatic fetch-rebase-push on rejected pushes).
 - **Undo never rewrites history.** Undo (⌘Z) and redo (⇧⌘Z) restore earlier states as **new forward commits** — never reset, never force. The whole trail stays inspectable in any git client. The one deliberate rewrite anywhere in the app is pull's rebase of **unpushed local** commits (07-sync-collab.md); published history is never touched.
@@ -48,7 +48,7 @@ The pathfinder's message engine carries over as the model — it is what earns t
 
 Switching a branch from the popover's picker (or creating-and-switching from the settings sheet — the 2026-07-31 split) — **create-and-switch runs the identical settle → flush → stamp → switch sequence, no at-HEAD fast path** (blessed 2026-07-31): "creating at HEAD can't change the tree" fails as a proof under concurrent writers (a foreign commit can land between the check and the create), and flushing pending work onto the branch it was made on is the honest cadence regardless — the occasional save-or-discard beat on a create buys one sequence with no special case:
 
-- **Settle the editors first — explicitly, never silently.** Branch switch neither silently commits nor silently abandons an open card-body Edit session: if any open card window has one (unsaved keystrokes, or on-disk ~700 ms saves the session hasn't committed — the mid-session state Auto-commit above deliberately leaves uncommitted), the switch presents a **save-or-discard step**: **Save All** ends every session with its normal commit (each card's Edit→Preview flip), **Discard** reverts buffers and uncommitted saves to HEAD, **Cancel** keeps the current branch and the sessions. **Open raw-source buffers are settled by the same step** (settled — an unsettled raw buffer is the worse hazard: its Apply later writes the *entire* pre-switch `index.md` byte-for-byte onto the new branch's card): Save All *applies* each raw buffer — and since Apply validates, a buffer that fails validation cancels the whole switch with focus on the offending window, nothing half-switched; Discard exits raw source without writing; Cancel keeps everything. External checkouts the app can't gate are the accepted last-writer-wins case, same as the Edit buffer (05-card-window.md's dirty-buffer rule; on git boards the overwritten version is a commit, one revert away). Silently flushing the commit alone would be wrong twice over: the tree can be clean precisely because a save hasn't landed, and a later debounced save would write old-branch text onto the new branch's card. With sessions settled, the pending auto-commit flushes (flush-before-overwrite above) and checkout runs on a truly settled tree: it cannot fail dirty, and no in-flight work is lost or dragged across branches. **The flush is Save-All-shaped** (blessed 2026-07-31): after Discard, the session's reverted bytes are *reconciled, never flushed* — the pending window for that folder drops, since disk again agrees with HEAD and committing the discarded saves would betray the button; pending changes elsewhere on the board still flush normally. (Inline title editors need no step of their own: reaching the popover's picker or the sheet's create control commits them — click-away commits, and board-scoped commands disable while one is focused — 04-interactions.md ▸ Grammar.)
+- **Settle the editors first — explicitly, never silently.** Branch switch neither silently commits nor silently abandons an open card-body Edit session: if any open card window has one (unsaved keystrokes, or on-disk ~700 ms saves the session hasn't committed — the mid-session state Auto-commit above deliberately leaves uncommitted), the switch presents a **save-or-discard step**: **Save All** ends every session with its normal commit (each card's Edit→Preview flip), **Discard** reverts buffers and uncommitted saves to HEAD, **Cancel** keeps the current branch and the sessions. **Open raw-source buffers are settled by the same step** (settled — an unsettled raw buffer is the worse hazard: its Apply later writes the *entire* pre-switch `index.md` byte-for-byte onto the new branch's card): Save All *applies* each raw buffer — and since Apply validates, a buffer that fails validation cancels the whole switch with focus on the offending window, nothing half-switched; Discard exits raw source without writing; Cancel keeps everything. External checkouts the app can't gate are the accepted last-writer-wins case, same as the Edit buffer (05-card-window.md's dirty-buffer rule; on git boards the overwritten version is a commit, one revert away). Silently flushing the commit alone would be wrong twice over: the tree can be clean precisely because a save hasn't landed, and a later debounced save would write old-branch text onto the new branch's card. With sessions settled, the pending auto-commit flushes (flush-before-overwrite above) and checkout runs on a truly settled tree: it cannot fail dirty, and no in-flight work is lost or dragged across branches. **The settle also clears each open card window's fine undo stack** (ruled 2026-07-31): pre-switch steps describe the branch being left — Save All and Discard alike end with every window's stack empty, the board-stack discard-and-reseed precedent one level down; the windows stay open, following their cards onto the new branch with fresh stacks. **The flush is Save-All-shaped** (blessed 2026-07-31): after Discard, the session's reverted bytes are *reconciled, never flushed* — the pending window for that folder drops, since disk again agrees with HEAD and committing the discarded saves would betray the button; pending changes elsewhere on the board still flush normally. (Inline title editors need no step of their own: reaching the popover's picker or the sheet's create control commits them — click-away commits, and board-scoped commands disable while one is focused — 04-interactions.md ▸ Grammar.)
 - **The undo/redo stack does not survive a switch.** It is discarded and reseeded from the new HEAD's first-parent ancestry — the relaunch rule applied at switch time; redo starts empty. (Replaying a restore commit from the previous branch onto the new one would be wrong.)
 - **Everything remote-facing tracks the current branch**: ahead/behind, Pull/Push, and push-on-commit all operate against the current branch's upstream. On a branch with no upstream yet, the first push — manual or push-on-commit — **creates it on the remote quietly** (`push -u` semantics): creating a remote branch is non-destructive, and quiet is consistent with push-failures-never-nag (07-sync-collab.md). Genuine failures queue with the badge as usual.
 - The switch itself is bracketed (02-architecture.md): watcher suspended, one full reload at the end. If that final reload fails, the board locks read-only until a successful reload — see 02's live-reload resilience; the on-screen snapshot is from the previous branch and must not be edited over the new one.
diff --git a/DESIGN/13-native-undo.md b/DESIGN/13-native-undo.md
index 9a350da..6c9dbe3 100644
--- a/DESIGN/13-native-undo.md
+++ b/DESIGN/13-native-undo.md
@@ -17,7 +17,7 @@ The undo/redo substrate for **every board without app-managed git, in every tier
 
 ⌫'s undo is the move back — a delete is a move into `.trash/` (cards resettled 2026-07-28; lanes rejoined 2026-07-29), so its undo is the ordinary inverse move, returning a card to its source lane and rank, a lane to its strip position (subtree intact — it never left the folder); a restore-by-move undoes the same way in reverse. The stack and the trash never conflict — they are the same folder moves addressed by recency instead of by selection. The old lane-delete recreate-from-capture inverse is **retired** — no destructive delete remains outside a trash, so nothing needs byte capture. A **permanent delete registers no step** — the trash's Delete and Empty Trash are not undoable (Rules above), lanes and their freight included; the confirm is the safety.
 
-**Comments keep the no-capture rule true — on the window stack** (re-ruled 2026-07-31, superseding the board-stack routing): a comment delete is a move into the card's `comments/.trash/` (01-storage-format.md ▸ Enhanced schema — the materialized-trash pattern one level down), its inverse the ordinary move back, and the step lives on the **card window's own stack** (Rules above) — the board stack never carries a granular comment step, so the old stale-after-close skip scenario cannot arise. **The purge of `comments/.trash/` defers with the coarse step** (re-ruled 2026-07-31, superseding purge-at-close): the coarse close step's undo restores deleted comments, so their backing lives as long as the step does — the purge runs when the coarse step leaves the board stack (undone-and-superseded, dropped off the end, or gone stale) or the board session ends; crash residue still sweeps at the next card-window open — **and residue is defined by the purge-deferral condition itself** (ruled 2026-07-31): `comments/.trash/` content referenced by a live coarse step on the board stack is a step's backing, not residue — the open-time sweep consults the stack and skips owned content, re-arming when the owning step leaves the stack (which is exactly when the deferred purge wanted to run; one condition, two consumers). Reopening a window can therefore never destroy its prior session's undo backing. Unowned content sweeps as before, armed-then-cleared like every heal memo. On Pro the substrate is history: the close commit nets delete-plus-purge to a removal, revert restores it, so purge rides the close flush there as before — purge timing follows the undo substrate's need.
+**Comments keep the no-capture rule true — on the window stack** (re-ruled 2026-07-31, superseding the board-stack routing): a comment delete is a move into the card's `comments/.trash/` (01-storage-format.md ▸ Enhanced schema — the materialized-trash pattern one level down), its inverse the ordinary move back, and the step lives on the **card window's own stack** (Rules above) — the board stack never carries a granular comment step, so the old stale-after-close skip scenario cannot arise. **The purge of `comments/.trash/` defers with the coarse step** (re-ruled 2026-07-31, superseding purge-at-close): the coarse close step's undo restores deleted comments, so their backing lives as long as the step does — the purge runs when the coarse step leaves the board stack **cleanly** — undone-and-superseded, or dropped off the end — or when the board session ends; **a stale-skipped step's backing instead survives to board-session end** (ruled 2026-07-31, decoupling skip from purge): the skip banner says nothing was applied, and an irreversible purge riding that gesture would be surprise loss — the skip is exactly when the user may want to inspect what the collision left; crash residue still sweeps at the next card-window open — **and residue is defined by the purge-deferral condition itself** (ruled 2026-07-31): `comments/.trash/` content referenced by a live coarse step on the board stack is a step's backing, not residue — the open-time sweep consults the stack and skips owned content, re-arming when the owning step leaves the stack (which is exactly when the deferred purge wanted to run; one condition, two consumers). Reopening a window can therefore never destroy its prior session's undo backing. Unowned content sweeps as before, armed-then-cleared like every heal memo. On Pro the substrate is history: the close commit nets delete-plus-purge to a removal, revert restores it, so purge rides the close flush there as before — purge timing follows the undo substrate's need.
 
 ## Out of scope