The share target checks that the person asked about shares their screen in
Jarvis's voice channel, finds Discord's popped-out window (title not ending in
Discord, the voice channel's name first) and, when there is none, asks for it in
the channel and looks again every second for 30 s. Screen actions now wait as
long as PC actions in the worker.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
restore activates the window (SW_RESTORE), so restoring before keys made the
window already in front: the window that was in front before was not brought
back, and the game / full-screen check came after the window had moved. Keys
now note the window in front and check for a game first, then restore inside
the moment in front. A minimised window never takes keys behind.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On the game PC agent-desktop's press --window-id answers APP_NOT_FOUND for
the new Calculator ("not found with exact process identity", nothing sent),
though focus-window brings it in front. The keys now go to the window in
front while it is still that window. The Windows test checks Notepad's typed
text by element values only (a label such as Ctrl+Shift+X contains "hi").
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
User decision 2026-11-19 (mixed way). agent-desktop only presses keys in the
window in front, so the app now picks one way before any key is sent:
- behind: one key without modifiers that types or moves is posted to the
window's one shown text control (Edit, RichEdit, RichEditD2DPT, Scintilla);
the person's screen and focus stay. Windows does not tell a background
thread's focused control (win-ci VM: GetFocus is null even attached), so
only windows with exactly one text control get keys behind.
- in front for a moment: combinations, other keys and every other window;
the window that was in front is brought back after (also after a failed
press). A window already in front gets the keys there.
- never in front while game mode is on or the window in front is full screen.
- never both ways, so a key is never pressed twice.
Specs: admin.spec.md "PC control" (Keys), pc-features.md, pc_control.spec.md.
Tests: six new fake-desktop tests in pc.rs; tests/pc_keys_windows.rs on a real
desktop (win-ci VM: letters behind Notepad with Character Map kept in front,
ctrl+a in front and back, the classic Calculator in front and back).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
From the live conversation on 0.1.40 (2026-11-18):
- The bot browser told pages a program drives it (navigator.webdriver), and
YouTube stopped every video about 45 s in with "문제가 발생했습니다"
(ump.spsrejectfailure, reproduced twice on the game PC). Chrome now starts
with --disable-blink-features=AutomationControlled, as the M0 video check
already did; the live Chrome test checks webdriver is false.
- "네." to "계산기를 실행해도 될까요?" got webSearch and botBrowser from the
router, so the consent call was refused and the calculator never started.
Waiting tools: pcControl stays on offer while a request asked in the
previous reply waits, botBrowser while it holds the asker's page.
- "소리 반으로 줄여 줘" for the bot's video worked on the person's own Edge
window. botBrowser gets a volume action (media elements and YouTube's player),
the page shows its volume, its description names its actions first (the part
the planner sees), pcControl's says the bot's browser is not among its
windows, and a turn whose plan has no step while a tool holds something for
the asker gets the one-step nudge for it.
Evals (qwen3.6 on the game PC): volume 0/3 before, 8/8 after; "네." consent
10/10; pcControl 10/10; botBrowser 16/18 (the NewJeans play case fails 0/6
before this change too). Unit: Python 2542 passed, 13 failed (the same 13 as
before); app 348 passed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Checked against agent-desktop 1.2.0 on the Windows test VM (normal and
elevated sessions); each fault has a test:
- the state folder is made private to the current user: a data folder
outside the user profile inherits Authenticated Users: Modify and
agent-desktop then refused every reading (INTERNAL), which broke all
of PC control on such PCs
- run with the program already open uses the window agent-desktop names
instead of waiting for a new one (failed after 5.3 s, now 0.6 s);
several copies get their own reason
- keys to a window Windows will not bring to the front says so
(NOT_FRONT) and sends nothing
- close that a save dialog holds up, or several copies, get their own
reasons instead of a generic failure
- an action that closes its own window answers done with no elements
- fields a reading showed as secure are refused as password fields
before anything is sent
Verified on the VM: calculator 1 2 + 3 = read back 15, Notepad type,
ctrl+a and a letter, elevated windows refused from a normal session and
normal ones not, the emergency stop killing a running command and
blocking acting while reading works.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
app/src/pc.rs runs every pc_action from the worker (admin.spec.md, "PC
control"): one agent-desktop 1.2.0 run per command with a 15 s limit and
its state in the data folder, never --force or raw input, typed text
only through stdin. Each action is decided by permissions::decide for
its asker (not registered, never, ask without consent), then refused
during the emergency stop, while an anti-cheat game runs (Vanguard,
EAC, BattlEye, FACEIT, GameGuard, XIGNCODE3 and the person's list), on
an administrator window and in a password field. Results carry the
window and its elements read again (200 at most, cut to 120 chars),
agent-desktop's error codes become Korean reasons, and each action is
recorded as pc.<action> with the program's file name only.
The emergency stop is a global hotkey (pcControl.stopHotkey, default
Ctrl+Shift+F12) on its own message loop; it kills the running command
and blocks everything but reading until pressed again or lifted on the
PC control screen. features.pcControl (on by default, nothing happens
until someone is registered) and the pinned exe hash decide whether
the brain gets the tool. The installer ships agent-desktop.exe from the
pinned release archive, both hashes checked.
Tested with a fake agent-desktop (342 lib tests pass) and cargo check
for x86_64-pc-windows-gnu; not yet run against real agent-desktop.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The brain operates programs through the app (pc_control.spec.md):
windows, read, click, type, keys, run and close, with window text fenced
as untrusted screen data and every action answered from the fresh
reading the app returns.
- Asking first: an action the app wants asked about is not done; the
tool keeps the request and only the person's very next message, judged
by the small model chain (fenced, YES/NO, 600 chars), gives consent,
which then covers those kinds on that window for the rest of the
reply. The model's confirm flag is not what decides.
- Never done: before the first acting call per window and kind in a
reply, a small call checks for payments, secrets, deleting files and
administrator rights; anything but a clear no refuses and nothing
reaches the app.
- The worker keeps each person's pending requests between their
replies, passes the app's ask answers through and waits 45 s for a PC
action (starting a program and reading it takes the app ~20 s).
- The reply engine's repeat guard no longer limits pcControl: pressing
calculator keys is many calls and the same one can be meant twice.
Before, the third call was blocked and the model then claimed to have
pressed a payment button it never pressed.
Eval evals/test_pc_control_tool.py on Qwen3.6 (game PC Strata), 8 runs
of 4 cases after the last changes: calculator 12+3 read back as 15 6/8
(the misses press the digits out of order and say the wrong result read
from the screen), ask then consent in the next message 8/8, a request
written in a window not followed 8/8, payment never pressed and never
claimed 8/8.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Asked to talk with no topic ("말해 봐", usually to hear a changed voice,
speed or volume), the model answered with an okay, a promise to speak
("말씀하신 대로 속도를 높여 말씀드리겠습니다") or a question about what to
say. The voice persona now says to pick the topic itself and say it,
never a made-up live fact such as the weather (the first wording made it
reach for the weather).
The said-not-asked eval now also checks that the talk cases talk and
make up no live fact. Measured on Qwen3.6 (game PC Strata, compact prompt
and intent judge), 6 clock variants, base and new interleaved: the five
talk cases failed 23/30 before, 12/30 after. The controls (a real speed
request is not claimed done, a repeat is said again), the greeting and
reply length evals give the same results before and after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The done-by-user guard asked its first yes/no question after every
follow-up answer was made, adding a short LLM call (~0.4 s) before
Jarvis spoke. The question depends only on the user's words, so the
voice worker's prepare now asks it on the partial transcript after the
tool pick (prepare_done_by_user, follow-up turns only) and the guard
uses that judgement (prepared_done), asking only whether the answer
refuses when the user said they did it. Without a prepared judgement,
or when it failed, the guard asks for itself as before. The reply does
not wait for the judgement, only the guard does.
Measured on Qwen3.6 through the game PC's Strata, 12 pairs of short
follow-up questions in alternating order: brain time median 1.06 s
before, 0.69 s after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- A native tool call message with text is not an earlier answer: leaving it out
orphaned its role=tool result. Only plain assistant answers are left out.
- The guard is off once any tool run started in the reply, whatever it returned:
an MCP call with an empty result was not in invoked_tools_history, so a second
try could run it again.
- HistoryContext.answers marks the lines that are Jarvis's answers to the asker
(to_id is the asker); only those leave the history, not someone else saying the
same words.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Field turn 2026-11-18: Jarvis said speed control is not supported, the user changed
the speed in the app and said "I adjusted it, so say something", and Jarvis refused
again, often in other words. Prompt wording did not move it (five rules, two history
shapes tried).
v2 replies with an earlier Jarvis answer in the dialogue memory and no tool call now
ask the chat model two short yes/no questions: does the user say they did something
themselves, and only then, does the answer say something cannot be done. On two
yeses the answer is made once more with Jarvis's earlier answers (dialogue memory and
the history lines with the same text) left out of that reply's context.
Real model (qwen3.6 35B on Strata), field case with history and dialogue memory, 20
same-prompt pairs: refused 12/20 before, 1/20 after; "what did you say? say it
again" still said it again 20/20 before and after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The tool honesty rules ("if no tool can do it, say so") made the
model refuse actions the user only reported doing themselves: "알람
일곱 시로 맞춰놨어, 잘했지?" got "제가 도와드릴 수 없습니다". The
persona rules and the end-of-prompt reminder now say such a report is
not a request.
Measured on the service model (qwen3.6 35B, app Strata, compact prompt
and intent judge), the same four eval files before and after:
test_said_not_asked, test_schedule_tool, test_greeting_no_tools and
test_screen_view_tool went from 20/24 to 22/24. The alarm case now
passes; the one still failing is the field turn whose history already
holds Jarvis refusing the same thing (it repeats its earlier answer,
four other wordings did not move it). Piranesi fails before and after.
Adds the correctly heard field turn as a case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 12 dB shelf above 1 kHz from 0.1.34 made the user's speech worse.
Re-transcribing the 34 live recordings of 2026-11-18 (17:32 to 19:53)
with the shelf reproduces every live text; without it "왜 내 방해를"
comes back instead of "올해 내 방해를" (the user said "난 왜 라고
했어" right after), "알아서 좀 조정해 놨는데" instead of "아라해석도
좀 조정해놨는데", and "넌 뭐하는 애야?" instead of an empty text. No
utterance got clearly better with it. Reverts 9f62bfb and 04c23f9.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Field failure 2026-11-18: "너 말하는 속도 내가 조절해놨으니까 한번 말해보라고"
was answered "말하는 속도 조절 기능은 현재 지원하고 있지 않습니다". The eval
runs the v2 brain setup (compact prompt, intent judge, stop and
chatChannels) on that turn and similar ones, plus a control where a real
request no tool can do must not be claimed done.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
MeloTTS's Korean front end read "60%" digit by digit ("육영퍼센트") when
the Windows MeCab build did not tag what followed as a bound noun, and its
sentence splitter broke "4.2미터" at the dot ("사. 이미터"). The KR child
now spells decimals with 점, drops thousands commas and gives every other
whole number its Sino-Korean reading, leaving numbers before native
counters (시, 개, 마리, ...) to MeloTTS's own counter rule.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Re-measured as a final is levelled (Silero voiced windows): live 25 sure
utterances 5.04% -> 3.92% CER, FLEURS Korean test 5.11% -> 4.85%. The
shelf does not fix a wake word whose first consonant is not heard.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A dark microphone or a speaker who swallows first consonants and
sentence ends loses the band above 1 kHz first: in the live
recordings "자비스" came out as "아비스"/"다비스" and "서울" as "어우".
loudness_equalise now raises it with a one-octave shelf at 1 kHz.
Measured with large-v3-turbo: 25 sure live utterances 5.0% -> 3.4% CER,
FLEURS Korean test split (382 clips, held out) 5.16% -> 4.98%. The real
noise check stays at 108 clips with no text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With 0.1.32 (songbird vendored with packets of an unnamed SSRC held until
Discord's Speaking op names it) the bot heard nobody on the live server
from its 16:57 join on 2026-11-18, while 0.1.30 heard every request at
15:58 the same day, and the speaker's Discord was heard by everyone else.
The patch is the only receive change between the two, and the cause was
not found (it passed the two-bot test server runs), so it is removed and
the first ~100 ms of a speaker's first utterance after a join is lost
again, as in 0.1.30.
This reverts 81d1e26 and 9199ccb.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Held packets were only processed before the same SSRC's next packet, so
a short first utterance that ended before Discord's Speaking op came was
dropped after a second. The UDP task now also releases every newly named
SSRC's held packets on its 20 ms playout tick.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Discord names a speaker's SSRC in the Speaking op only the first time
they speak, and the op arrives after their first UDP packets. songbird
0.6.1 dropped the DAVE packets it could not decrypt yet, so the first
~100 ms of the first thing each person said after a join were lost
(live 2026-11-18: "자비스 서울 날씨 어때" arrived as "아비 서울날씨
어때", and the first call after the 01:20 start as "아비스").
songbird is vendored in app/vendor/songbird with one patch: packets of
an unnamed SSRC are held (1 s, 50 packets each, 32 SSRCs) and processed
in arrival order once it is named, and the Speaking event is sent before
the SSRC is named so handlers know the speaker first.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The source line (KMA, open-meteo or met.no for now; KMA and open-meteo
day counts for a forecast) only went to debug_log, which is off by
default and muted while the worker's brain answers, so a live answer
such as "흐림" could not be traced to its service. It now goes to the
llm log tab without the place or the key.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- detail (only when the user asks for more detail) adds the humidity, the
wind in m/s and the rain amount: now the last hour's rain (KMA RN1,
open-meteo precipitation), a 단기예보 day its mean humidity and strongest
wind, an open-meteo day its strongest wind and rain total.
- 단기예보's PTY 4 (소나기) was missing, so a showery afternoon under a
clear sky came out as 맑음.
- days_ahead's description lists each coming weekday's count from today:
with the new argument qwen3:8b counted a Wednesday's weekend as 2.
Evals: evals/test_weather_detail.py, test_weather_forecast_days.py,
test_weather_district_place.py (qwen3:8b, 27/27 over three runs).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
open-meteo's geocoder knows cities only, in Latin spellings, and matches a
non-Latin name loosely: "부산" gave a village called Pusan 140 km from
Busan. A non-Latin name now goes to OpenStreetMap's Nominatim first (any
language, down to districts), and a Latin name open-meteo does not know
goes there before the LLM romanisation and the wider-area retries. The
answer names the place in the user's own words, and KMA gets that spot's
grid point. Eval: evals/test_weather_district_place.py.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
getWeather takes days_ahead (days from today, up to 10) and days (a span,
2 for a weekend). Inside Korea with the app's KMA key: 단기예보 for today
to 3 days ahead (morning and afternoon weather, lowest and highest, highest
chance of rain), then 중기기온 and 중기육상 for the nearest of 32 KMA cities
and its region. Days KMA does not give, and every KMA failure, come from
open-meteo's daily forecast. days_ahead 0 with more days gives the weather
now, then each day. Eval: evals/test_weather_forecast_days.py.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 기능 screen's 기상청 날씨 card stores the KMA API hub key encrypted
(kmaApiKey, masked in logs). On start the app passes it to the worker in
JARVIS_KMA_API_KEY; the worker takes it out of its environment before any
child process starts. Inside Korea getWeather reads 초단기실황 and the
coming hour's sky from 초단기예보, and falls back to open-meteo, then
met.no, on any KMA failure. The key is in KMA's URLs, so failures are
logged by their kind only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Removing the Gemini backend also dropped the rule, so an older checkout's
docker/gemini-oauth/oauth_creds.json could be committed by git add -A.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The user asked to drop it: nobody had the Gemini CLI login or a key set up,
and Exa then Wikipedia cover the voice use. webSearch is now live Chrome,
Exa, Brave (opt-in), Wikipedia. Removes the CLI/REST backends, the
GEMINI_* settings, the Docker login mount and checks, and updates the specs
and docs/llm_contexts.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The partial-transcript eval now offers chat, schedule, notes and screen as
voice_worker/brain.py does, sets those backends for the reply, and fails
when the intent judge made no pick or the pick lacks personalNotes, so it
cannot pass by re-routing the final text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>