The enforcement rules must not depend on any harness's skill/trigger
mechanism — the auto-loaded entry point itself now instructs every
agent to read agent/VERIFICATION.md before its first change (and
THEME-DESIGN.md / TESTING.md where relevant), instead of merely
pointing at them.
Verification: V0 (docs-only).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The repo is maintained by agents from multiple vendors, so agent
instructions move out of vendor-specific locations into shared,
git-tracked markdown:
- AGENTS.md is the new entry point for any harness; CLAUDE.md becomes a
symlink to it (Claude Code keeps working unchanged).
- Skill bodies relocate to agent/: VERIFICATION.md (the enforcement
rules, ex .claude/skills/nomarchy), DELEGATION.md (capability tiers
light/standard/frontier, scout/runner role contracts, token economy,
parallel fan-out — consolidates the CLAUDE.md model table, LOOP.md's
economy section, and skill §6.5 into one place; vendor model names
survive only in the per-harness mapping table), THEME-DESIGN.md
(ex .claude/skills/theme, which previously lacked frontmatter).
- .claude/ shrinks to a thin Claude Code adapter: settings, subagent
defs, and skill shims that route into agent/.
- Maps updated: agent/README.md (instructions vs state vs adapters),
docs/README.md, README.md layout tree.
Cleanup: stale old_distro .gitignore entry dropped (dir long gone);
local result*/__pycache__ artifacts removed; JOURNAL.md rotated (29
recent entries kept, 120 older moved to agent/JOURNAL-ARCHIVE.md,
rotation rule documented in the header).
Verification: V0 (docs/meta only) — nix flake check --no-build exit 0
with the tracked symlink; grep sweep confirms no vendor model names
outside DELEGATION.md's mapping table.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Push #76's swapfile-resume residual into a VM as far as headlessly possible.
A full hibernate→resume from a RUNTIME swapfile turned out not to be
headless-reachable: systemd reports CanHibernate=no for a swap that isn't
declared in the boot config, and a runtime-computed resume_offset can't be
fed into the boot cmdline (the genuine chicken-and-egg). So the check now
verifies, bounded and deterministically, what the partition test
(checks.hibernate) can't and what a VM CAN prove: the installer's btrfs
@swap swapfile is built correctly and its offset is valid — NOCOW
mkswapfile, file-type swapon, a good `map-swapfile -r` offset the kernel
accepts (/sys/power/resume_offset), and zram (prio 100) above it.
The literal encrypted-swapfile power-cycle + LUKS-initrd unlock stays a V3
laptop check (the dev box runs exactly this layout, /proc/cmdline verified).
(First cut had an unbounded wait_for_shutdown that hung on the CanHibernate
refusal; rewritten with a thread-bounded wait + a CanHibernate pre-check,
which gave the clean verdict. Device is /dev/vdb without useNixStoreImage.)
Verification: V2 PASS (nix build .#checks.x86_64-linux.hibernate-swapfile,
fresh, exit 0); flake check --no-build green. #76 residual narrowed in
BACKLOG/ROADMAP.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two low-severity edge bugs from a fresh-eyes review of the session batch:
- cmd_auto: an inverted schedule (sunrise > sunset, i.e. a day window that
crosses midnight) stayed perpetually on the day theme — the old
`not between` shortcut is an empty interval when sunrise > sunset.
Replaced with a real wrap-around: daytime = mins >= sunrise or
mins < sunset, so both directions switch. Normal sunrise < sunset path
is unchanged (checks.auto-theme still holds).
- rofi autotheme HH:MM validator: `[0-2][0-9]` accepted 24:00–29:59.
Tightened the case pattern to `[01][0-9]:[0-5][0-9]|2[0-3]:[0-5][0-9]`
so out-of-range hours are rejected (they'd otherwise skew the boundary).
(Two other review findings dismissed: papirus DOES ship breeze/breeze-dark
so that iconPacks prefix is correct; the doctor RAM-unknown row is cosmetic
since /proc/meminfo is always readable.)
Verification: py_compile; boundary truth-table (normal + wrap-around);
HH:MM case table (23:00/19:59/20:00 accept, 24/25/30:00 reject); auto
--which normal-case unchanged; menu bash -n; flake check green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add an "Auto theme (day/night)" subsection to docs/OVERRIDES.md § 1
(Appearance), after the Icon pack section. Covers what it does (light/dark
switch on a schedule, same one engine), how to turn it on (Look & Feel ›
Auto theme, or the shell set/auto sequence), the settings.autoTheme
fields, and behavior (periodic ~15-min re-check, idempotent rebuilds,
instant disable). Names/fields/flags verified against rofi.nix,
nomarchy-theme-sync.py, and the icon section's style.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Bernardo rightly pushed back on deferring #79 to a real session: a
simulated sunset IS VM-testable. Add checks.auto-theme, a runNixOSTest
that installs nomarchy-theme-sync, points NOMARCHY_DEFAULT_THEMES at the
presets, stubs the rebuild via NOMARCHY_REBUILD (a marker), configures
the day/night pair + schedule, then WALKS THE VM CLOCK across the
boundary: 19:55 -> 20:05 -> 20:06 -> next-day 07:30, asserting the active
theme goes summer-day -> night (1 rebuild) -> night (idempotent, still 1)
-> day (2 rebuilds).
This isolates the auto decision + apply + rebuild-trigger; the actual
home-manager switch is stubbed (that generic apply path is exercised by
every manual theme change). The only on-hardware residual is the user
timer firing on its OnCalendar schedule, which systemd-analyze validates.
#79 drops its [blocked:hw] tag — it's V2-verified. ROADMAP updated;
lesson recorded in MEMORY (don't default timer/session features to V3:
stub the generic step, simulate the clock).
Verification: V2 PASS (nix build .#checks.x86_64-linux.auto-theme, fresh,
exit 0); V0 flake check green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds an "Auto theme (on/off)" row to Look & Feel and an `autotheme`
submenu: toggle the feature, pick the day/night themes (from
`nomarchy-theme-sync list`), and set the sunrise/sunset times (HH:MM,
validated). This is the menu surface over settings.autoTheme, completing
the auto time-of-day theme feature (slices: CLI / timer / menu).
Enabling writes the flag and defaults the summer-day/summer-night pair
--no-switch, then runs `auto --force`, so a single rebuild both installs
the timer and applies the correct theme for now. Disabling is instant
(--no-switch): the flag flips and `auto` self-gates, so the lingering
timer no-ops until the next rebuild. Day/night/time writes are live
--no-switch (read by `auto` at tick time).
Verification: V1 — HM generation builds (bash -n on nomarchy-menu at
build) + bash -n on the built menu. Logic test with stubbed
rofi/theme-sync/notify: selecting the toggle from off runs exactly `set
day/night/enable ... --no-switch` then `auto --force` — the intended
one-rebuild path. V0 flake check green. All three #79 slices shipped;
the live timer switch across sunset is a real-session (V3) check on a
Nomarchy machine.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
modules/home/autotheme.nix installs a user service + timer
(nomarchy-auto-theme) that periodically runs `nomarchy-theme-sync auto`,
so the day/night theme is applied on the current schedule through the
same one engine. Install gates on settings.autoTheme.enable (read from
state, the way updates.nix gates on its enable), so a machine not using
the feature carries no timer.
Timer: OnStartupSec=1min (settle after login) + OnCalendar=*:0/15 +
Persistent (catch a transition missed while off). Periodic rather than
exact sunrise/sunset timers — robust to suspend/DST, and `auto` is
idempotent (no rebuild unless the theme must change) so ticks are cheap.
Verification: V1 — real-config eval (state temporarily enabled, then
restored) confirms the units land when enabled (OnCalendar *:0/15,
ExecStart `... nomarchy-theme-sync auto`) and are absent when disabled;
systemd-analyze validates the schedule. V0 flake check green. The live
timer -> home-manager switch on a real Nomarchy session is pending a
Nomarchy machine (same blocker as #76).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First slice of the auto time-of-day theme pair (VISION § D, split into
CLI / timer / menu). `nomarchy-theme-sync auto` reads
settings.autoTheme.{enable,day,night,sunrise,sunset}, picks day vs night
from the local clock, and applies the matching preset — but only if it
differs from the current one, so the slice-2 timer can fire freely
without triggering needless rebuilds. `--which` prints the decision
without switching; `--force` and `--no-switch` included.
cmd_apply's core is factored into apply_named() so `auto` drives the
exact same one engine (no second pipeline). apply behavior is unchanged.
Verification: V1 (package builds) + functional tests on the built binary
against scratch state — wide-day window → day slug, a 00:00–00:01 day
span → night, disabled → no-op, missing day/night → clear die, auto
--no-switch writes state, a re-run no-ops as "already on", and plain
apply still writes (refactor intact). V0 py_compile + flake check +
checks.theme-sync-validate green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add docs/OMARCHY.md for users coming from Omarchy (VISION § F). Leads
with the one real difference — imperative dotfiles vs a declarative flake
built via the menu, with rollback through NixOS generations — then maps
the deltas: install (ISO/nomarchy-install or MIGRATION.md), a familiar
keybindings table, theming (SUPER+T / nomarchy-theme-sync), installing
software (home.packages / nix shell, no AUR), and update+rollback
(nomarchy-pull/-rebuild/-home, boot-menu generations, nomarchy-doctor).
Registered in docs/README.md; VISION § F marked shipped.
The Nomarchy side is grounded in the code: keybindings pulled from
keybinds.nix, and every cited command (7), the nsearch alias, and the
named theme slugs (gruvbox/rose-pine/nord) verified to exist. Omarchy-side
claims kept to known facts (Arch, dotfiles, pacman/AUR).
Verification: V0 (docs-only); commands/slugs/binds checked against the tree.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
nomarchy-doctor gains a read-only hibernate/sleep section (the highest-ROI
Day-2 gap, a natural #76 follow-on): resume device on the kernel cmdline
(+ resume_offset for a swapfile), disk swap >= RAM, zram active, and any
suspend/hibernate error in the previous boot's kernel log. Self-gates
(skip on swap=0) and emits only ok/warn/skip, so it never fails the sheet.
Adds `hibernate` to the checks.doctor probe list so the row is asserted.
Fixes a set -euo pipefail hazard the first VM run caught (exit 1, no
verdict): a no-match grep inside $(...) and a `cond && action` both abort
under set -e — now guarded with || true / || echo 0 / if. Recorded in
MEMORY as a reusable gotcha for writeShellScriptBin scripts.
Verification: V2 PASS — checks.doctor VM test green (fresh build, exit 0;
asserts the hibernate row renders, healthy run stays 0). V1 — ran on real
LUKS+@swap hardware: all three ok rows render (resume set, swap 34G >= RAM,
zram 16G) and the sheet reaches its verdict. V0 bash -n + flake check.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
QA/drift sweep: deterministic checks + bash -n/py_compile all pass
(the 4 "needs-args" scripts pass via their flake derivations); drift
recon found none (dead code, stale refs, TODOs, changed-area
contradictions, counts all clean — 24 presets/previews consistent);
eval sweep green. No fixes needed.
VISION-gap pass: verified candidates against the tree first, which
killed two false gaps (nomarchy-rebuild/home already tail logs + point
at doctor on failure = §A; nvd diff = §B; Firmware menu already exists).
Filed three verified-open PROPOSED items: doctor hibernate/sleep section
(§C, natural #76 follow-on), Omarchy migrant one-pager (§F), auto
time-of-day theme pair (§D, big). Await human triage.
Verification: V0 (agent-notes/docs only); flake check --no-build green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The only doc for the icon-pack override was a comment in the downstream
template. Add an "Icon pack" subsection to docs/OVERRIDES.md § Appearance
(the user-facing appearance-customization doc): the
`nomarchy-theme-sync set icons "<name>"` command, why only Papirus ships
by default (~1 GiB; extra packs opt-in), a known-packs table, and that
the choice is sticky across `apply` (presets carry no icon field).
Example names verified by building all four packs and listing their real
theme directories — corrected Qogir to its capitalized Qogir-Dark/Light
(a lowercase name silently falls back to hicolor). Hibernation needs no
new doc (README swapfile note + MIGRATION.md runbook + Power-menu row).
Verification: V0 (docs-only); command + variant names checked live.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
theme.nix now resolves nomarchy.theme.iconThemePackage from a small
icon-pack registry (Papirus/Tela/Qogir/Reversal/Numix, matched by name
prefix). Default themes name Papirus-*, so the resolved package is the
plain papirus-icon-theme derivation — byte-identical closure, zero added
MB. Only when a theme's `icons` names a set from another pack is that
pack (and only it) symlinkJoined alongside papirus. stylix.nix consumes
it via mkDefault. Adding a pack is one iconPacks row.
Deliberately NOT shipping more packs by default: papirus alone is already
~1 GiB and each extra pack adds ~hundreds of MB, so variety is opt-in per
install (set `icons` in theme-state.json), not a default closure cost.
Documented in templates/downstream/home.nix + ROADMAP § Icon themes.
Verification: V1 (HM activationPackage builds). Eval proofs — default
resolves to plain papirus (zero delta); icons="Tela-dark" ->
nomarchy-icon-themes symlinkJoin (papirus+tela) with the name flowing,
without building tela; all 5 pack attrs exist in the pinned nixpkgs;
template-sot + option-docs + flake check green. No default visual change
(same papirus), so no screenshot; the opt-in render is downstream.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The dev box runs Bernardo's own NixOS (TuringMachine, 25.11), not
Nomarchy — so hibernating it tests that config's hibernate, not
Nomarchy's oom.nix/patch-template/rofi wiring, and can't close#76's V3.
Its layout (LUKS cryptroot + btrfs @ + /swap/swapfile + resume_offset)
does resume, which corroborates the MIGRATION.md runbook, but nothing
more. Rescope the two #76 HARDWARE-QUEUE entries to require a machine
actually running Nomarchy (a Latitude install, or the dev box after
migration), and fix the matching reference in the BACKLOG #76 stub.
Verification: V0 (docs/agent-notes only).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add checks.hibernate: a runNixOSTest modeled on nixpkgs' hibernate.nix,
wired to the actual risk #76 introduces — coexistence with the zram swap
that oom.nix now enables by default. It imports oom.nix, puts the disk
swap at a lower priority than zram with an explicit boot.resumeDevice,
and uses systemd-initrd. The test drives a full
hibernate -> power-off -> resume -> crash cycle: a volatile ramfs marker
survives resume but NOT a crash-boot (proving genuine resume), and it
asserts zram sits at priority 100 while the disk swap is lower, so the
hibernate image lands on disk rather than volatile zram.
What QEMU genuinely can't do — feed a boot-time resume_offset for a btrfs
swapfile (offset only exists after first-boot allocation) or drive a
LUKS-initrd unlock — stays a V3 laptop check (HARDWARE-QUEUE, dev box
already runs that exact layout), plus the no-swap notify in a real
session. With this, all V0/V1/V2 work on #76 is done; #76 collapses to a
[blocked:hw] stub and the design record moves to ROADMAP.
Verification: V2 PASS — nix build .#checks.x86_64-linux.hibernate built
the driver fresh and exited 0. nix flake check --no-build green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Per the settled #76 UX call: keep the Power-menu Hibernate row
unconditional, but when hibernate fails (typically a swap=0 machine with
no disk swap), surface a notification pointing at the enable-hibernation
runbook instead of a silent no-op. On success `systemctl hibernate`
blocks until resume, so the notify only fires on a real failure.
Verification: V1 — homeConfigurations.nomarchy.activationPackage builds
(rebuilds nomarchy-menu; writeShellScriptBin runs bash -n). Control-flow
proved on the arm extracted from the built script: failing systemctl ->
notify fires, succeeding -> no notify. flake check --no-build green.
V2 hibernate->resume VM test + V3 laptop remain open (#76 slice 3).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a self-contained "Enabling hibernation on an existing machine (no
reinstall)" section to docs/MIGRATION.md, per the settled #76 call to
document migration rather than ship a tool. Covers: @swap subvolume
creation (subvolid=5), btrfs mkswapfile (NOCOW), reading the resume
offset via map-swapfile, and wiring fileSystems."/swap" + swapDevices +
resumeDevice + resume_offset into system.nix. Flags the /swap mount as
required on the hand-edit path (a fresh install inherits it from
disko-generated hardware-config). Notes the zram-priority reservation,
a no-LUKS variant, and the swap=0 opt-out.
Verification: V0. All read-only commands run live against this machine's
LUKS(cryptroot)+btrfs(@)+/swap/swapfile layout; nix flake check
--no-build green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Bernardo's three design calls on the hibernation half of #76:
- swap sizing = exactly RAM (installer default, no change)
- migration for existing machines = docs runbook (MIGRATION.md), not a tool
- no-swap Hibernate = keep menu row, notify on failure
Recon confirmed new installs are already hibernation-ready (installer +
patch-template resume wiring + Power-menu Hibernate row). Folded decisions
into BACKLOG #76; re-sliced remainder to MIGRATION.md runbook + notify-on-
failure menu; recorded in Decisions.
Verification: V0 (docs-only) nix flake check --no-build green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First slice of BACKLOG #76 (hibernation + zram, split): zramSwap on by
default in modules/nixos/oom.nix — the memory-pressure layer before
earlyoom. zstd @ 50% RAM (nixpkgs defaults, explicit for reproducibility),
priority 100 so day-to-day paging fills zram first and any disk swapfile
stays reserved for #76's hibernate image (hibernation can't resume from
volatile zram). Follows the earlyoom mkDefault precedent — no toggle.
Adds checks.zram-swap runNixOSTest importing oom.nix, asserting the device
is an active swap, uses [zstd], and carries priority 100.
Verified: V2 checks.zram-swap PASS (/dev/zram0 in /proc/swaps, zstd,
prio 100); V1 full distro toplevel builds; V0 flake check --no-build green.
#76 remainder (disk swapfile default + resume + installer) still open.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Bernardo verified the GTK file-chooser (OpenFile) portal floats on real
hardware; the xdg-desktop-portal-gtk float/center rule already shipped in
hyprland.nix. Drop the "OpenFile not re-probed / optional residual"
caveats in HARDWARE-QUEUE and journal the V3 close. Bookkeeping only, no
code change.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Agent must V2-test hibernation in the VM (encrypted BTRFS @swap
swapfile). Reference layout from live Latitude: no zram yet; swap
inside LUKS. MEMORY pointer for implementers.
Extract nomarchy-pull/rebuild/home to pkgs/nomarchy-lifecycle; install
via system + home. Home profile wins on PATH so home-manager switch
installs the fixed pull (skip git when no upstream) without waiting
for a system rebuild that needed a working pull first.
V0: flake check --no-build green.
charge_control_end_threshold was written (80) but charge_types stayed
[Adaptive], which ignores the cap while still reporting it — battery
kept charging past 80% on Latitude. Oneshot now selects Custom when a
limit is set, Adaptive/Standard when off; sets start hysteresis; doctor
warns if type ≠ Custom while limited.
V0: flake check --no-build green. V3: re-apply limit, confirm
charge_types shows [Custom] and status stops Charging above cap.
Class xdg-desktop-portal-gtk (+ Nix .…-wrapped) from package desktop/
libexec. Polkit already V3 PASS on Latitude. NEXT only #20 remains.
V0: flake check --no-build green.
Two bare `{:%…}` specs emptied the hover tooltip (box with no text).
One strftime block with an embedded newline shows date + zone; ensure
tooltip label colors on generated + summer bars.
Machine flakes often have no remote (local auto-commits only). Distro
updates come from nix flake update of the nomarchy input. Only git pull
--ff-only when @{u} is set; always update the lock; print nomarchy rev.
dmenu mode is case-sensitive by default (dmenu compat), so config.rasi
case-sensitive=false fixed apps/drun but not nomarchy-menu pickers.
rofi_menu helper always passes -i; networkmanager_dmenu gets rofi -i.
V0: flake check --no-build green. V3: type sys in root Menu → System.
Align summer-day/night waybar tokens with JSON roles (hierarchy split
on night; day overlay≠muted), unify bar font 13px + battery 25/10, add
VPN pill and status states (.on/.available/.recording/…). Recapture
executive-slate and neon-glass picker previews from theme-shot.
Verified: V0 contrast+wholeswap; V2 theme-shot four slugs.
