Several sessions can now share one phone: a lease covers one package (the
real screen stays the default scope, so existing callers keep their
meaning), and only phone-wide actions take `--whole-phone`, which is capped
at 300 s and runs its `--cleanup` commands exactly once on release or, via a
detached reaper, on expiry. The serial is always resolved, never the
"default" placeholder that let android_ui and phone_deploy lease different
keys for the same phone (2026-09-25). Adds status --json, holds and whoami
for phone_vd.sh and the raw-adb hook.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012Ajzz98xksbfak8CaunZWC
Filled scalloped canopy over a stroked J-handle (filled because a stroked
dome with ribs reads as a jellyfish at 48dp). AppIcon moves to _app_icon.py,
re-exported from apps.py, so the registry stays under the 250-line cap.
Prunes the two oldest evidence artifacts (60-file cap; still in history).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012Ajzz98xksbfak8CaunZWC
Two sessions driving the same phone interleave their taps: on 2026-09-20 a
tap meant for the workout sandbox opened another session's manga reader,
twice. The lease is a flock-guarded JSON file per serial under
~/.cache/phone-lease/, owned by the nearest ancestor `claude` process (one per
session), refreshed on every android_ui call and expiring 180 s after the
last one, so a quiet session frees the phone without anyone remembering to.
A foreign fresh lease makes a call wait up to 60 s, then exit 3 naming the
holder and pointing at the alternatives (the phone's virtual display, the
headless emulator) instead of inviting a retry loop. phone_deploy.sh takes
the same lease (exit 40).
The oldest evidence artifact is pruned to stay under the 60-file cap; it
remains in history.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YbBZqUaNWyhHYfzz6zGC7q
`secret-tool lookup` is non-interactive — on a locked keyring it returns
nothing and shows no dialog — and a busctl Unlock+Prompt pair cannot work
either, because Secret Service prompt objects die with the D-Bus
connection that created them. secretstorage keeps one connection:
get_default_collection().unlock() pops the gcr dialog and blocks until it
is answered. Verified on the real profile: 5 v11 rows -> v10, then with
the keyring re-locked Helium (through the wrapper, so the flags file and
LeechBlock apply) lists all 5 logins and keeps every row.
`pgrep -f /opt/helium-browser-bin` also matched the invoking shell when
its argv mentioned the path, so the guard refused on an idle machine;
now anchored to `^/opt/helium-browser-bin/`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M8YGXVZdNUG79q5DGsdMm3
Chromium 151 removed the last Manifest V2 code paths and the Web Store
purged uBlock Origin on 2026-08-31, so ungoogled-chromium 153 cannot load
the full blocker any more. Helium (GPL-3, ungoogled-chromium base) bundles
uBO 1.75.0 as a component extension and keeps MV2 alive.
- installers/install_helium.sh: pacman deps, AUR via yay or makepkg
(+validpgpkeys import), /usr/local/bin/helium-browser ->
browser-preexec-wrapper so LeechBlock NG and the hosts refresh still
apply, --clone-profile with a downgrade check and running-browser
refusal, then the verifier. Passed in vmbox from scratch, then host.
- installers/verify_helium.py: Xvfb + DevTools exit-code gate — uBO
background_page + manifest_version 2, every /etc/chromium/policies key
on chrome://policy (Helium reads that dir natively), adblock-tester >= 95.
- installers/migrate_helium_passwords_to_basic.py: lightdm autologin
leaves gnome-keyring locked, so the 5 v11 logins make Helium prompt on
every launch; this re-encrypts them as v10 and only then enables
--password-store=basic.
- register helium-browser in aur_packages.txt and the periodic browser
wrapper/service lists; pylint hook gets websockets + cryptography;
codespell learns the vendor name; anthropic 1.6.0 -> 1.7.0 (gate red);
oldest evidence artifact pruned to stay under the 60-file cap.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M8YGXVZdNUG79q5DGsdMm3
kuhy: "I do not want it to work like this, I want it to just work
normally." PS4 mode (046D:C260) cannot be supported on this PC (no kernel
driver, SDL lg4ff has no c260 either, so FFB is dead there too) and the
selector is physical. Keep the mode detection so the journal and --status
name PS4 mode instead of "not found"; drop the desktop notification and
the 5-minute Restart loop.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NVfC2T6DyvCFqse9A99zef
On the 2026-09-15 and 2026-09-17 boots the wheel enumerated as 046D:C260
(PS3/PS4 selector in PS4 position, bound by hid-generic). The BPF udev
rule, the ensure unit and its hotplug rule all keyed on C24F, so the unit
failed once with "no G29 found" and nothing was said; 1st/3rd stayed on
the firmware's bad decode until the selector was flipped a day later.
- fixes/g29_wheel_mode.sh (new, sourced): classify the HID bus as
native/ps4/compat/absent; root-safe desktop notification via
/run/user/*/bus + sudo -u.
- --ensure: absent -> exit 0; ps4/compat -> exit 1 + critical
notification; native -> attach. --status prints the mode first.
- g29-shifter-bpf.service: Restart=on-failure, RestartSec=5min so a
notification lost before the session bus exists at boot is re-sent.
- 99-g29-shifter-ensure.rules: also fires on C260.
- tests/test_fix_g29_shifter.bats: 13 tests against a fake sysfs tree.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NVfC2T6DyvCFqse9A99zef
The dkms pacman hook already rebuilds every preset on install and remove,
so the unconditional mkinitcpio -P cost 20 s twice per run on the host.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SS8M5wk88nJTXxDWz61KKk
Live on this host since 2026-09-11 but never committed. The wheel firmware
decodes the shifter's hall sensor itself and this unit's top row rests at
Y 180-193, below the ~195 engage threshold, so 1st/3rd often did not
register. g29_shifter.bpf.c re-decodes those two gates from the raw X/Y
bytes with hysteresis (engage >= 160, release <= 140) and ORs the bit in
before hid-logitech parses the report; it never clears a bit, so the
firmware's working gates cannot regress. fix_g29_shifter.sh builds,
attaches and installs it persistently (udev-hid-bpf rule + a systemd
safety-net unit pulled in on every re-enumeration); g29_shifter_capture.py
records and grades raw captures per gate.
Two files were over the 250-line cap: the test file is split at the
record/CLI boundary with the report builders moved into a conftest fixture
(the naming hook allows only test_*.py in tests/), and the grader's
docstrings are tightened. No behaviour change to the installed fix.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SS8M5wk88nJTXxDWz61KKk
The "flags nothing on a quiet machine" case read the real /proc/loadavg and
went red whenever a game was running on the host (22/16 during a BeamNG
session), and the "flags a load at/above thread count" case had to SKIP
whenever the host was idle. Same override pattern as JOURNALD_CONF_DIR: the
harness writes a fixture, both cases now assert deterministically, and a
missing file degrades to 0 as before.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SS8M5wk88nJTXxDWz61KKk
BeamNG.drive 0.39 native lost G29 FFB because SDL3's HIDAPI lg4ff driver
advertises an effect-status query it cannot answer (12,201 failures per
session in beamng.log). DOCS-g29-ffb.md records the diagnosis, the narrow
SDL_JOYSTICK_HIDAPI_LG4FF=0 hint vs the broad one, and the launch-option
%command% bug that had gamemoderun passed as a game argument.
install_new_lg4ff.sh swaps the in-tree hid-logitech for new-lg4ff via DKMS
(dkms.conf installs it under the same name in updates/dkms/), regenerates
initramfs presets that bundle the module, and offers --status/--revert.
Passed install/status/revert in vmbox; host run pending.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SS8M5wk88nJTXxDWz61KKk
300M held about two days on kuhy's desktop: one Steam boot wrote 480k
journal lines and the evidence for a curfew failure two weeks earlier was
already gone when it was needed (2026-09-13). The three Arch copies of
the journal tweak (arch_hardware, arch_perf_fixes, arch_perf_report) move
to a 4G cap, and "large"/vacuum now mean past that cap — a 2G journal is
deliberate retention, not a finding. Tests follow.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015mLnUeoJ3mUwGRTowTUH6d
A push took ~13s. prettier is a pre-push hook precisely because Node's
startup dominates its runtime, and pre-commit chunks a large file list into
several invocations -- a 106-file push paid ten cold Node starts. ci-mirror,
the hook that looks expensive, was already 0.02s once its green-tree cache
hit.
A file whose content, prettier version and prettier config are all unchanged
since its last clean check cannot have become unformatted, so the verdict is
now cached and the file never reaches Node. Only clean verdicts are recorded,
so a formatting regression is always re-checked; the key includes
`prettier --version` and every config file, so an upgrade or a config edit
invalidates the cache wholesale. The systemd-run memory cap is unchanged.
Measured on the same push range: 13.0s -> 3.3s total, of which ~0.6s is
hooks and the rest is git's own network negotiation. On a 40-file list the
runner itself goes 0.652s -> 0.057s, starting no prettier process at all.
endurain/README.md and copilot-instructions.md are reflowed: the ~
reorganisation lengthened paths inside them, which changed prettier's
markdown table alignment.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AwtowrowDXx6ccmWVPkCaF
Also bumps the 16 meta/requirements pins the freshness gate flagged
(anthropic 1.5.0, numpy 2.5.3, pytest-randomly 5, flake8-bugbear 26.9.9,
ruff 0.16.7, pyright 1.1.414, selenium 4.49.0, hypothesis 6.168.0, lxml
6.1.3, pipdeptree 4.2.5, the two types-* stubs) and Flutter 3.47.3 in
meta/.fvmrc. One test still expected the pre-move ~/mcp-servers path.
The pre-commit hook fed .zsh files to shfmt and shellcheck, neither of
which parses zsh (SC2148/SC1071), so no zsh file could be committed since
the hook was tightened on 2026-08-22; those now get `zsh -n` instead. The
hook's own shfmt pass also staged its reformat of 8 shell files.
CI_GREEN_SKIP because main is red on exactly the dependency-freshness
finding this commit fixes, which is the documented use of that flag.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AwtowrowDXx6ccmWVPkCaF
`verify` gated on the two invariants whose breach cut every bridged
container off the internet on 2026-08-29 and again after the 2026-08-30
reboot -- but nothing invoked it, so it reported drift only when a human
thought to ask, and the outage it exists for is exactly the one nobody
thinks to ask about.
nftables-drift.service now runs it after nftables.service, docker.service
and network-online.target at every boot, plus a daily Persistent timer. A
breach is a failed unit in `systemctl --failed`, not a log line: a
self-check that detects a failure has to gate, not warn.
Proven in ~/utils/vmbox across two real reboots -- drifted /etc leaves the
unit failed with the diff and the missing docker0 accept in its journal,
matching /etc leaves it active with no false positive -- then installed and
run green on the host. Not yet observed across a host reboot.
Also bumps the gatelock pin to v0.7.3, which is what turned main's
dependency-freshness job red; CI_GREEN_SKIP is set because this commit is
that fix.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015d48t4qsQLBJnDwDeo2Nom
wake_alarm._alarm no longer exists: the PC's waking role was deleted in
kuhyx/wake-alarm, and the phone owns the alarm now. This removes ALARM_MODULE,
_run_alarm() and the --with-alarm flag, leaving one leg -- the workout lock.
Landing this in the same change as the deletion is the point. An orchestrator
left pointing at a module that no longer exists is exactly the failure this
repo has already had once: after screen-locker was extracted on 2026-05-28 the
reference here went stale and the workout lock died with ModuleNotFoundError on
every run for a month, while the alarm half kept working and hid it.
morning-routine.service loses --with-alarm and its comments now name what
actually starts it: wake-alarm-trigger.timer, at the phone's synced alarm time,
not the deleted systemd-sleep hook. Environment=DISPLAY=:0 stays -- the comment
credited it to the alarm's Tk window, but workout-locker.service needs it too.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PcdngP688mW6rJye3LRkSN
The forward-chain fix from 334c0cd9 only ever reached the live ruleset by
hand; /etc/nftables.conf kept the bare default-drop chain, so the next
reboot cut every bridged container off again and the Signal bot went deaf
a second time. Nothing compared generated against persisted.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The GTK tray icon was the only status surface, and it had been commented out
of ~/.config/i3/config since 2026-08-21. With nothing in the bar, the locker
miscounted the week for a fortnight and the machine never said so.
Shows "<count>/<minimum> · <gaming budget>", read from the local status API
(screen-locker-web.service on 127.0.0.1:8770); clicking opens the full UI.
Three things it deliberately will not do quietly:
- a dead status server renders "status server down" in red, or the last known
value marked (stale) in orange -- never a blank block. That server is what
the gaming budget reads, so its absence costs two hours;
- a disarmed or unscheduled timer overrides the compliance colour entirely
and shows "DISARMED" in red. A locker that cannot fire is the worse
problem and the one that hides;
- clicking only opens the UI. There is no bar action that can clear a lock.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Fcwu78wk17FWMHbbfbyicc
`write_nftables_ruleset` emitted a forward chain with `policy drop` and no
rules at all. That chain runs at the same netfilter hook as Docker's own
DOCKER-USER/DOCKER-FORWARD chains, so it vetoed them: every bridged
container lost outbound networking while the host itself stayed online, so
nothing looked broken.
It cost the Signal bot three silent hours today. signal-cli lost its
connection to Signal at 15:32:50 and looped on reconnect for 2.5h, while
the bot kept answering /healthz with 200 -- `/v1/groups` still served from
local state, so every surface said healthy.
`flush ruleset` also deletes Docker's own tables, and dockerd only rebuilds
them on its next network event, so restore_docker_rules() now restarts
docker right after a successful apply.
Verified in vmbox (Arch guest, docker installed, alpine probe):
BASELINE=reachable -> OLD=blocked -> NEW=reachable, sshd active throughout.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KbjfWdAyHbCubac7wJKJZW
A new app at ~/restaurant-rater rates restaurant dishes, so it needs a
family icon. shield-cutlery is diet-guard's and the accent is shared
fleet-wide, so the new glyph has to be distinguished by silhouette alone: a
filled star above an open bowl -- a dish, and a verdict on it.
The star is filled rather than stroked for the same reason as the nib and
the anvil: at 72px weight its five interior notches close up well above
MIN_NEGATIVE_SPACE and it renders as a pentagon at 48dp.
The bowl went through two passes. The first used shallow walls stopping at
y=736 and read unmistakably as a smiling mouth with the star for a nose --
two curves with a shape above them is a face unless the vessel is clearly
deeper than it is wide at the base. The committed version drops to y=752
between narrower walls, which reads as a bowl.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012YDbT3ip2yeT18W3KCZdLq
Two failure modes compose cannot see, both of which report a healthy
container while it serves 404s:
1. The Caddyfile is a bind mount, so its contents are not in the config
hash compose compares. Editing it changes nothing observable, compose
says Running, and the new directives never load.
2. A bind mount resolves once, at start. Replacing dist/ rather than
emptying it leaves the container on the old unlinked inode: /srv reads
empty and every request 404s.
Hit both at once on a real deploy today -- the container was Up 6 hours,
Caddy logs clean, /srv empty, curl 404. Reproduced by rm -rf dist (inode
84695030 -> 84695031) and fixed by the patched function in one run.
wait_for_website also stops using curl -sf, which collapses every non-2xx
into the same "did not become reachable". It now reports the status it saw,
because unreachable and answering-404 are different bugs and the message
named only the first.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A7BaCSV8FEtcNgVDVsRAJj
The personal-website repo now prerenders each post to
dist/blog/<slug>/index.html. Caddy try_files does not resolve a directory
to its index on its own, so with the single-term directive every post URL
fell through to the SPA shell: right content once React booted, wrong
<title>, and no link preview. Verified against caddy:2.8 in Docker both
ways before and after.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A7BaCSV8FEtcNgVDVsRAJj
The guard prints `CI_GREEN_SKIP=1 git commit ...` and stops there, but it runs
at two stages: at commit, and again inside the pre-push CI mirror's
clean-worktree `pre-commit` run, which sees the same still-red baseline. So the
documented escape gets you past the commit and into an identical wall at push,
with no hint and nothing in the output connecting the two.
Hit today landing the pre-commit cache-key fix: the commit went through, the
push did not. The variable is read from the environment and `ci_mirror.sh` does
not scrub it, so `CI_GREEN_SKIP=1 git push` carries through to the nested run --
which is now what the message says.
Verified both paths: with the variable set the script warns and exits 0, and
against today's green main it exits 0 silently. shellcheck clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SfnJmxxsZ534AKk16rC2Gr
`requires-python` and the CI pin already said 3.14, but ruff, mypy and pylint
were still grading `meta/` against 3.10 -- three tools judging an interpreter
nothing runs any more, and ruff withholding the pyupgrade fixes that only
become safe at the newer target. `pin_ci_python.py` rewrites these three knobs.
What the stale target was hiding: `datetime.timezone.utc` where `datetime.UTC`
is now correct (UP017), and ruff's stable formatter applying PEP 758 --
`except (A, B):` becomes `except A, B:`. That makes the source 3.14-only at the
syntax level, which is what `requires-python = ">=3.14"` already promised; kuhy
chose it deliberately over holding ruff a release back (2026-08-29).
Also drops a comment in `wsg_grabber/models.py` that the bump makes false: it
explained why the file used ``str, Enum`` instead of ``StrEnum`` under mypy at
3.10, sitting directly above a ``class Verdict(StrEnum)`` that had already
stopped obeying it.
ruff, ruff-format, mypy --strict and pylint all clean at the new target.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SfnJmxxsZ534AKk16rC2Gr
Counting transcript records reports 1.00 tool calls per message no matter
what the model did: the harness writes one JSONL record per tool_use block,
all sharing message.id. A hand-rolled per-record count read the real 1.9%
batching rate as 0.0%, which nearly shipped as a headline finding.
Group by message.id instead, and report it, so the number comes from a
tested code path with the window handled by find_transcripts rather than an
ad-hoc timestamp. Three snapshot fields give next week's run a baseline.
Axes was at the 12-attribute cap, so the counters live in their own Batching
dataclass with the two ratios as properties.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NiSSDu4gKbp9GGxTTtXvYW
track-bars on the family charcoal/accent: rising bars with a plotted trend
point, which reads as odds movement.
linux:false because betting-sim's desktop target is a Chrome-wrapped web build
rather than the GTK embedder, so the hicolor set would go unread.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RDSQS61Vgb6wptLPT2Utbf
A Flutter/Material TextField publishes its label as `hint` and leaves text
and content-desc empty, so every field on such a form arrived as an
anonymous EditText. The workout app's manual-workout form is nine of them:
the only way to fill it was by screen position, which is precisely the
mis-tap-writes-the-wrong-field failure this tool exists to prevent.
`hint` ranks below text and content-desc but above resource-id. Material
clears the hint once content arrives, so a filled field still answers to its
real text and a caller can tell an empty field from a completed one.
The app-side alternative does not work and is documented as a trap: wrapping
the field in Semantics(label:) emits a second labelled node at the same rect,
and MergeSemantics collapses them by dropping the hint from Android's tree,
which leaves the form MORE anonymous than before. Widget tests cannot catch
that -- they assert on Flutter's SemanticsNode while this tool reads
Android's AccessibilityNodeInfo, so nine of them passed green on the build
whose on-device labels had just gone empty.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AWq9MRemX9ZWvNmnhNWUP2
`Pre-commit checks` has been red on main since 2026-08-24: the md-naming
and file-length hooks are thin shims that exec the real gate out of the
utils repo, and a runner has no ~/utils, so both died with "shared gate
not found at /home/runner/utils/...". The dedicated md-naming.yml
already checks utils out into .utils/; this does the same and points
UTILS_ROOT at it, which is the override the shims already support.
CI_GREEN_SKIP is set because this commit IS that fix.
Also registers punchme (github.com/kuhyx/punchme) in the shared icon
family on the family accent, Android-only so linux=False. Its glyph is a
time card with a clipped corner, a column of four punched holes and
three ruled lines; the first draft put two holes above two wide slots
and read unmistakably as a face at every size.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
focus_owner (com.kuhy.focus_owner, the Device Owner enforcer) was the one
Flutter app never registered with the shared generator. It shipped a
hand-written vector adaptive icon added purely to unblock its build — the app
had no icon resource at all and every release AND debug build failed on AAPT
"resource mipmap/ic_launcher not found". That placeholder used #E8EAED, so it
was off-family as well as unregistered.
None of the existing 13 glyphs fits a device-policy enforcer, so this adds
padlock-closed: shackle stroked (inheriting the group's round caps), body
filled in the accent, keyhole punched back out in the field colour. A fully
stroked padlock loses its keyhole at 48dp and a fully filled one loses the
shackle's opening; splitting them keeps both.
Geometry was computed against style.py's constraints rather than eyeballed,
per the lesson recorded on _NOTE_ASCENDER: body 292..732 x 520..792, shackle
centreline r=132 about (512,520) so its outer top is 352 (inside the 232 safe
edge) and its legs land at 380/644, well within the body. The keyhole leaves
60px of body above it and 76px below the stem, both clear of
MIN_NEGATIVE_SPACE=36 — checked by rendering at 48px, where the hole survives.
The repo path points into ~/phone-focus-mode/focus-owner: focus-owner was
absorbed as a subtree on 2026-08-24 and both the standalone clone and its
GitHub repo are gone.
app_icons stays at 100% branch coverage; 44 tests pass.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KU9czAaz7wAkjyJ5uEB3t
Renames this repo's markdown onto README.md / CLAUDE*.md / DOCS*.md /
TODO*.md so a future session can tell a live task from a finished
record without reading the file to find out.
TODO files carry the literal "REMOVE ME AFTER FINISH" and are deleted
when the work lands. Enforced by the md-naming pre-commit hook and CI
workflow, both delegating to the shared gate in kuhyx/utils so this
repo's idea of a valid name cannot drift from every other repo's.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A47aX5DJwmA3QFn3pkEWi7
The previous commit made hosts-blocker's install.sh register the guard-lib
instances itself, which is what finally gives a fresh machine real hosts
enforcement. It also created a class of machine that --rollback refused to
touch: the state dir it demanded, /var/lib/guard-lib-migration, is written by
the one-shot migration, and a machine that never had a legacy layer has no
such state. So --rollback exited 1 while the instances it exists to remove
were live -- every freshly installed machine had no escape hatch at all.
Rollback now proceeds whenever a guard-lib instance is registered, says
plainly that there is no legacy layer to restore, and still refuses when
there is genuinely nothing to roll back.
The second half is worse and was only visible by running it. The existing
re-assertion of chattr +i works by starting hosts-guard.service and
hosts-bind-mount.service, neither of which exists on a fresh machine, so
every systemctl cat skipped and the rollback finished with /etc/hosts holding
no immutable flag and no bind mount -- less protected than before it ran.
It now re-asserts immutability directly when no legacy unit did it.
Verified in a pristine vmbox guest whose guards came from
install_core_system.sh --all: dry-run and real rollback both exit 0, the
three instances are uninstalled, shutdown-schedule (a different feature) is
untouched, and /etc/hosts keeps ----i----------------- and all 178410 lines.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
Two of the seven modules invoked python_pkg/ paths that stopped existing when
screen_locker and steam_backlog_enforcer were extracted into their own repos,
so the first thing the installer did on a fresh machine was fail. Nothing
installed guard-lib either, which meant the CORE "Midnight shutdown timer"
module always died with "guardctl not found on PATH".
Every module now resolves through install_module_from_repo, and
require_extracted_repo clones a missing sibling instead of only printing
instructions -- otherwise "clone it yourself" is not actionable on the bare
machine this script exists to set up. The clone drops privileges to the
invoking human via sudo -u, because an installer that re-execs under sudo
would otherwise leave a root-owned checkout in the user's home; a sandbox
cannot catch that, since the guest is root throughout. A clone that fails is
still a loud, fatal module failure, never a silent skip.
guard-lib is addressed as the "utils" repo: it is a subdirectory of
github.com/kuhyx/utils, so asking for "guard-lib" would resolve to
~/guard-lib and derive a clone URL that does not exist. It installs first,
before the modules that need guardctl.
Separately, the security hardening suite's screen_locker section hardcoded
$HOME/testsAndMisc/python_pkg/screen_locker/screen_lock.py. That path went
away with the extraction, so all five checks silently SKIPPED -- reporting a
clean run for tests that never executed, which is the exact failure mode
lib/extracted_repos.sh was written to prevent. It resolves the repo properly
now ($HOME is also wrong under sudo), and the assertions are rewritten against
the current implementation: the four symbols they grepped for no longer exist
anywhere, having been replaced by SPORT_CHOICES, SPORT_TABLE_TENNIS,
_validate_table_tennis and MANUAL_WORKOUT_REFLECTION_MIN_CHARS.
On this host that takes the suite from 18 passed / 1 failed / 5 skipped to
23 passed / 0 skipped. The task file's "18 passed / 6 skipped out of 17"
target was stale -- the suite is 24 checks now -- so it could not be matched
literally.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
Two post-push audits, both worth keeping.
vendor/common.sh sources three sibling libs unconditionally and only common.sh
had been vendored, so every entry point printed three "No such file or
directory" errors. Every suite passed anyway -- the functions in use are
defined above that block and a failed sibling source is non-fatal -- which is
exactly why it needed its own test. Fixed in digital-wellbeing c5eb028.
A grep across /etc, /usr/local and /usr/bin, broader than report_live_state.sh
(which reads only /usr/local/{bin,lib}, units and symlinks), found three
references. None from this work: pw_guard.sh's target still exists, and
dwm-session carries two pre-existing single_use/ paths. dwm is dormant and was
deliberately abandoned, so that one is reported rather than changed.
Also records that digital_wellbeing/data/ was safe to delete: never tracked on
any historical path, and no script in the unit references data/ at all.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
The last unit over the cap is now github.com/kuhyx/digital-wellbeing, cloned
at ~/digital-wellbeing, and `--warn` is gone from the directory-depth-cap
hook. `check_directory_depth.sh --all` exits 0 with zero violations, down from
46 at the start of the session. That was the point of the campaign.
History was filtered across all three paths the unit has lived at (63
commits); the current path alone holds 2.
The six suites that test this code lived in linux_configuration/tests/ rather
than beside it, so they were filtered separately (24 commits) and merged in as
a second root. A path filter of the unit alone would have left them here,
testing code that had left the repo -- and their names would still have been
in shell-tests.yml's hardcoded arrays, failing CI on missing files.
linux_configuration/lib/common.sh is vendored into vendor/ rather than trimmed
to the four functions actually used: set_actual_user_vars calls two of the
others, so a trim breaks silently. It is in vendor/ and not lib/ because
test_music_parallelism.sh copies the unit's own lib/ into a scratch worktree
while stubbing the shared lib; in one directory the real common.sh overwrote
the stub, and the suite silently measured the wrong library.
Running the callers caught what shellcheck and grep both missed:
check_and_enable_services.sh died with "DW_REPO: unbound variable" because the
definition sat below its first use. It now runs clean and reports ok for every
extracted-unit check.
Two long-standing live defects fixed as a consequence. The pacman-wrapper
drift manifest embeds ABSOLUTE paths and had named deleted scripts/ paths
since the flatten, so the check reported a permanent false "stale or
tampered"; it now verifies, with unchanged hashes proving the extracted bytes
are identical. And the last two stale live references are gone: 8 -> 0.
The two pre-commit excludes for setup_midnight_shutdown.sh are dropped rather
than moved. lsattr shows nothing in the unit is chattr +i any more, and the
fixer hooks leave the file alone when allowed to run.
Removing 80 shell files also shrank the jscpd denominator enough that the
virtualenv/tqdm clones inside the GENERATED .ci-mirror-venv/ tipped the repo
to 2.09% and blocked this commit, which adds no duplication. That directory is
the same vendored third-party shell the ignore list already excludes as
.venv/, just under another name; jscpd walks the working tree rather than the
index, so being untracked did not spare it. Excluded by name: 1.74%.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
161 tracked files, 3.2M. All 158 episodic entries have retrieval_count 0 --
the store was written to and never once read from. Last write 2026-05-28, the
tool is uninstalled, and nothing references it: the only code hit for "hippo"
anywhere in the tree is the blocked domain "wordhippo" in
leechblock_defaults.json.
No .gitignore entry was needed after all. The handoff anticipated one for
hippo.db, which was untracked only because *.db is ignored -- but with the
whole directory gone there is nothing left to ignore.
Tracked files: 1346 -> 1175.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
Five live mechanisms were silently pointing at paths that no longer exist.
None of them announced it; each degraded to a log line or a missing file.
- hosts-file-monitor.service is ACTIVE and watches /etc/hosts for tampering,
but its embedded installer path was the pre-flatten scripts/ one. It logged
"Hosts install script not found" 37 times: it detected tampering it could
not repair.
- dns-blocklist-refresh.service (enabled, next run 00:01) named a deleted
single_use/features/ path, as did media-organizer.service under utils/.
- setup_periodic_system.sh read all TEN of its templates from
periodic_background/system-maintenance/, which has held nothing but
untracked .pyc and .coverage residue since system-maintenance was
extracted. Every install of a service, timer or script was reading a file
that was not there.
- /etc/sudoers.d/hosts-install-no-passwd granted NOPASSWD on the old hosts
installer path.
Running the fixed dns unit then caught a bug the static checks missed: under
systemd there is no SUDO_USER and `id -un` is root, so the new resolver
returned /root/hosts-blocker and the service died with "Feed generator not
found". extracted_repo_home() now handles the bare-root case by falling back
to the owner of the lib file itself. Verified across all four invocation
modes: as the user, via sudo, as root with SUDO_USER unset, and with an
explicit *_DIR override.
The service now reports "Feed refreshed and dnsmasq reloaded (189038 lines)".
/etc/hosts is untouched throughout -- still 176888 lines, still chattr +i,
still bind-mounted by its guard.
Live artifacts were patched via validated temp files, never in-place: the
sudoers drop-in was checked with `visudo -cf` before install, and every
original is backed up under /var/tmp/*.pre-{extract,fix}.
Stale live references: 8 -> 2. The 2 that remain name digital_wellbeing
paths and can only be fixed when that unit is extracted.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
linux_configuration/periodic_background/hosts/ is now
github.com/kuhyx/hosts-blocker, cloned at ~/hosts-blocker. Clears 5 of the 46
directory-depth violations; the remaining 41 are all digital_wellbeing, so
--warn stays on the depth hook until that unit moves too.
History was filtered across ALL FOUR paths the unit has lived at (hosts/,
linux_configuration/hosts/, linux_configuration/scripts/periodic_background/
hosts/, and the current one). Filtering on the current path alone would have
preserved 1 commit out of 51 -- the flatten moved it recently, so a naive
single --path silently truncates the history to the flatten commit and still
looks fine in `git log | head`.
No source line changed: all 15 files are byte-identical to the copies removed
here. install.sh already resolved its lib/ via BASH_SOURCE.
The six callers now resolve the checkout through a new shared helper,
linux_configuration/lib/extracted_repos.sh, which reads the invoking user's
home from SUDO_USER's passwd entry rather than $HOME -- several of these
scripts re-exec under sudo, where $HOME is /root. A missing checkout fails
loudly with a clone instruction; it is never a silent skip, which is the
failure mode that left screen-locker's self-repair dead for five months after
the 2026-05-15 reorg.
The live blocker is untouched: /etc/hosts keeps its chattr +i, and the guards
load their plugins from /usr/local/share/guard-lib-plugins/ (byte-identical by
md5), not from a repo path.
Verified: 25 tests pass from ~/hosts-blocker; shellcheck clean on the helper
and all six callers; sudo resolution prints /home/kuhy/hosts-blocker;
test_setup_dns_blocker.sh passes; report_live_state.sh --check is byte-identical
before and after, so no new stale references.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
Shell tests went red on my last push. test_network_diagnostics.bats sourced
its scripts through $REPO_DIR/scripts/diagnostics, a level the flatten
removed, so setup() failed and all 30 tests errored.
Latent rather than new: the previous run was green only because ci-mirror
aborted before the bats stage. Fixing the concurrency OOM let the gate reach
this suite for the first time since the flatten.
Grepped the tree for the removed scripts/ and single_use/ levels; this was
the last one in tracked code.
Also rewrites the two handoff docs around the units that remain.
CI_GREEN_SKIP=1 because this commit is the fix for the red baseline.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
The push kept failing with "1798 passed" immediately followed by "Killed".
The suite was never the problem: it peaks at ~0.29 GiB against its own 2G
cap. ci_mirror.sh runs pre-commit --all-files and the pytest gate
concurrently, and the pre-commit tree peaks at ~1.75 GiB (mypy and pylint
repo-wide, invisible to /usr/bin/time because it measures only the parent).
The pair wants ~2 GiB at once; the pytest scope has MemorySwapMax=0, so on
a loaded machine it is the one that dies.
Concurrency is now conditional on MemAvailable. Above 4 GiB the fast path
is unchanged. The probe lives in ci_mirror_mem.sh, following the existing
ci_mirror_range/ci_mirror_shell split, which also keeps ci_mirror.sh under
the line cap.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
The pre-push gate was failing on a layout that is actually correct. The
script sources ../../lib/common.sh, which is right for where it sits after
the flattening; the test's guard still hardcoded ../lib/common.sh from
before the move, so it reported drift against itself.
The prefix is now read out of the script under test, so the two cannot
desynchronise the next time the tree moves. Verified non-vacuous: rewriting
the script to ../../../lib/common.sh in a throwaway worktree still fails
the guard, and it now names the real prefix.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
A gate that blocks a commit with no tooling to clear it just moves the
problem to whoever hits it. --prune stages the removal of the oldest
artifacts, keeping the newest 20 per directory plus the template, and
leaves the commit to the operator.
Running it caught a bug that reading it would not: sorting on %cs (date
only) ties every artifact committed the same day, so sort fell back to the
path and the prune deleted alphabetically - it picked one of today's
newest files instead of the oldest. Sorts on %ct with -rn now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
The prune commit's evidence noted that "a recurring prune or a hard cap
would be the durable fix if this recurs" - which is advice, not a fix. The
tree would have regrown at one artifact per commit and been noticed months
later, exactly as it was this time.
check_repo_size.sh now caps evidence/ and contracts/ at 60 files each: the
headroom the prune bought, against 21 today. main accumulates failures
rather than returning on the first, so tripping both the source ceiling and
the artifact cap reports both.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
These accumulate roughly one per commit and were the largest single driver
of tracked-file growth: 695 artifacts, 487 of them from 2026-08 alone.
Kept the 20 most recent of each plus both templates. A count cap rather
than a time window, because the growth is bursty - a 7-day window would
still have kept 393 files and bounded nothing. Selection is by git commit
date, not by filename: the corpus mixes -2026-05, -20260604 and undated
slugs, so a filename filter would have mis-sorted ~48 files.
Safe to delete: nothing reads the historical corpus. Every reference in
meta/scripts is either template.json or validation of files staged in the
current commit, and the three hooks still pass. sessions/ is append-only
and untouched.
Tracked files 1993 -> 1340.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
The depth cap had 104 violations. 57 are structural: whole projects with
their own layout (C/, dwm/, artgate/), a tool split into modules to satisfy
the line cap, prose trees, and directories whose name an external format
dictates (systemd, hooks, workflows, patches, fixtures). Those are exempted
the same way lib/ and tests/ already were.
The other 47 all sit under periodic_background/{hosts,digital_wellbeing},
which can only come down by extracting those units - blocked on a manual
guard repoint. So the hook runs --warn: the number stays visible on every
commit without blocking unrelated work. Enforcement is one flag away.
Also guards the argument order: `--all --warn` used to ignore --warn and
exit 1, which as a hook entry would have blocked every commit.
check_repo_size.sh is the companion gate - depth stops one subtree growing
a taxonomy, this stops the repo absorbing everything. It counts tracked
files (git ls-files, never find: the tree holds ~30k untracked extraction
leftovers) minus vendored and hook-generated trees: 1049 against a 1200
ceiling, printed every run so the threshold is auditable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
check_file_length.sh required a dot in the filename, so every
extensionless executable was silently uncapped. is_source_file() now
reads the first two bytes and accepts a shebang, which catches real
scripts while still skipping data files like i3/config.
That immediately surfaced boot_recovery/boot-repair at 756 lines --
three times the largest capped shell file in the repo, uncapped purely
because its name has no dot in it. Split along its own section markers
into lib/{discovery,repair_fs,repair_kernel,modes}.sh, leaving the entry
script at 218.
The report block stays inline on purpose. It is top-level code with exit
calls rather than function definitions, and sourcing it changed control
flow: 30 of 42 tests failed. That is recorded in the lib headers so the
next reader does not retry it.
No vmbox pass for this one, deliberately. boot-repair is UEFI/ESP tooling
and the guest is BIOS/GRUB, so a sandbox green would be actively
misleading -- the vmbox skill says so itself. The evidence is the fixture
suite, which drives the real script against synthetic root trees: 42
passed before, 42 passed after.
verify_shell_split.sh reports 31 of 32 functions byte-identical. The one
difference is clean_shadow_files, where quoting $ESP_MOUNTPOINT fixed a
genuine SC2086 that only became visible once the function left the
monolith -- shellcheck could no longer see the readonly declaration and
stopped assuming the expansion was safe.
lib/tests/run_all.sh delegates to that same suite rather than standing up
a second, weaker harness: these libraries need the entry script's strict
mode and globals, so testing them in isolation would test a fiction.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
i3 config was filed under periodic_background, which it never was. It is
now linux_configuration/i3/ (config + installer) and
linux_configuration/i3blocks/ (the status scripts), both at depth 2.
misc/testsAndMisc-bash/ was a directory named after the repo it already
lived in, so its contents moved up to misc/. Its stray .gitignore came
along, keeping the media rules working from the new location; its
.vscode/tasks.json did not, because it invoked Bash/transcribe.sh, a path
this repo dropped long ago.
Most of this commit is the references the previous flatten missed. Its
regex skipped anything with a variable prefix, which left 16 files --
including check_and_enable_services.sh and install_core_system.sh, the
orchestrators -- still naming $CONFIG_DIR/scripts/periodic_background/...
It had also rewritten three i3 paths into scripts/i3blocks/, a directory
that has never existed; the i3blocks suites caught that by failing.
13 more symlinks broke on the i3 move. Repointed, and cpu_monitor/volume/
disk were run again to confirm the bar still reports.
Depth 139 -> 104. Roughly 47 of the remainder sit in digital_wellbeing
and hosts, still deferred. The rest encode real structure -- a C project,
Python package directories, artgate's routes/styles, systemd unit dirs --
so flattening them further would destroy information rather than
taxonomy.
Verified: 1798 tests at 100%, both i3blocks suites pass, the shell
coverage gate is clean, and the relocated media rules still ignore what
they should while test_fw.srt stays tracked.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
