docs(agent): Latitude sweep #2 booked — 6 passes, SUPER+? bug filed (item 26)

PASS: waybar supervisor resilience, snapshots polkit+GUI (restore/
rollback residual queued), sys-rebuild, night-light on/off,
auto-commit, hibernate single-unlock (1b0eeea confirmed).
Partial: auto-timezone (enabled; zone check pending).

Item 26: SUPER+? never fires — `$mod, question` can't match because a
real ? carries SHIFT in the modmask; fix is `$mod SHIFT, question`.
SUPER+SHIFT+K no-op was the comma-layout gate working (none
configured) — retest annotated in HARDWARE-QUEUE.

PROPOSED: menu parity for downstream-flake options (Bernardo) — every
hand-edit-the-flake option gets a menu writer, per the in-flake-state
philosophy.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

agent/BACKLOG.md

@@ -22,9 +22,21 @@ next, in what order*.
 
 ## NOW
 
-*Items 21–25: real-hardware QA findings from the Latitude 5410
+*Items 21–26: real-hardware QA findings from the Latitude 5410
 (Bernardo, 2026-07-04) — bugs found on metal outrank the queue.*
 
+### 26. SUPER+? cheatsheet bind never fires on hardware
+Reported (Latitude sweep #2): SUPER+? does nothing — the ? reaches the
+focused terminal. Cause: keybinds.nix binds `$mod, question`, but
+typing ? holds Shift, so Hyprland sees mods SUPER+SHIFT with keysym
+`question` and its exact-modmask match skips a bind declaring SUPER
+only (the "question already implies Shift" comment doesn't hold on
+real input; headless V1/V2 never pressed the key). Fix: `$mod SHIFT,
+question` (user-facing label stays SUPER+?); while there, audit the
+bind list for other shifted keysyms (none today — letters carry
+explicit SHIFT mods). **Verify:** V1 (rendered bind string in the HM
+generation); V3 retest on the Latitude.
+
 ### 25. swaync text unreadable on summer-day (light-theme contrast)
 Reported: on summer-day, notification/control-centre text is invisible
 (contrast). Root cause (traced 2026-07-04): swaync.nix renders `.body`
@@ -211,6 +223,22 @@ earns ≥3 entries — keep the root at six.
 *Agents: append here with a one-paragraph pitch (what/why/cost). Do not
 implement. Bernardo moves accepted items into a tier.*
 
+- **Menu parity for downstream-flake options** (raised by Bernardo,
+  2026-07-04, from the hardware sweep): anything the docs or the
+  hardware queue ask a user to *hand-edit into the downstream flake*
+  (`keyboard.layout` multi-layout, `power.batteryChargeLimit`,
+  `updates.enable`, `services.printing`, …) should have a menu toggle
+  instead — the in-flake-state philosophy already gives the menu a
+  write-into-the-flake path, and night-light / auto-timezone /
+  auto-commit prove the pattern (incl. rebuild-on-first-enable).
+  Pitch: (a) an audit listing every user-facing `nomarchy.*` option vs
+  its menu surface; (b) menu writers for the gaps, respecting the
+  toggle-vs-package discipline (bare template packages stay hand-edits)
+  and menu-placement convention (right submenu, root stays short).
+  Values beyond booleans (layout lists, charge thresholds) need a small
+  input UI per option. Cost: audit is one iteration; the writers a few
+  more, sliceable per option.
+
 - **Portal/Flatpak camera picker still lists the internal IR sensor**
   (ROADMAP § Webcam follow-up). The shipped IR-hide is a *WirePlumber
   v4l2* rule, but Flatpak/portal apps consume cameras via the