The game logs into HORIZON_HOME and falls back to the player's own game folder when it is unset. Harnesses
that set only HORIZON_NO_AUDIO (and test_native_client, which set nothing) wrote logs, and on a world run
advancements, into ~/Library/Application Support/horizon.
lab.game_env gives every game a fresh HORIZON_HOME, HORIZON_LAUNCHER_HOME, XDG_DATA_HOME and XDG_CONFIG_HOME,
no audio and no launcher self-update; a folder the caller passes is kept. Every launcher of the game uses it,
parity.py gives each harness its own HORIZON_HOME, and tests/test_game_home.py fails for a script in tools/
that starts the game without it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
Horizon Settings gains a row: "AI Agent..." and "No AI: Off". No AI hides the
AI Agent page and closes the control socket at once.
The AI Agent page has an Agent Control switch (horizon-extras.txt
agentControl, off by default): on, the game the player starts opens the
default control endpoint, so horizon-ctl and MCP clients reach it without
--control. Turning it off closes the endpoint at once; dropping a control
server now also stops its listener thread. The page shows whether an agent
is connected, when it last sent a command and the socket, and Copy
Connection Details copies the socket path, a horizon-ctl line to try and the
MCP config for this game.
agent_control_e2e.py --scenario setting covers it: Agent Control on opens the
endpoint without --control, the pages list their widgets, Copy works, and
turning the switch off removes the endpoint while the game keeps running.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The table gains what only the raw socket could do (options pages, connect
and invite codes, disconnect, LAN, Play with Friends, books, signs, mount
inventories, effects, xp, enchant, potion, fill, containers) and new
commands: worlds, world-rename, world-backup, world-delete --confirm (the
Select and Edit World screens, from the title screen only), servers (the
Multiplayer list) and interact-entity (right-click an entity). status,
mouse, pad, release, bench and stats become MCP tools; summon takes a
position. Destructive and network tools carry MCP annotations.
docs/horizon-ctl-commands.md is generated from the table
(horizon-ctl help --markdown) and a test fails when it drifts. The
agent_control_e2e manage scenario exercises the new commands.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
horizon_core::no_ai reads and writes the noAi line of the game's
horizon-extras.txt, so the launcher and the game share one switch. While
it is on the game never opens its control socket, horizon-ctl refuses every
command with exit 4 and kind denied, horizon-ctl mcp and horizon-launcher
mcp refuse to start (and refuse calls if it is turned on mid-session), and
horizon-launcher no-ai [on|off] shows or sets it. The game's extras keep
the key when they rewrite the file. Both e2e harnesses check it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: merged after verify passed. telemetry.playhorizon.dev is live on Railway (service telemetry in horizon-site, volume /data): /healthz, /v1/live (a test ping counted) and /v1/benchmarks/summary (CORS for playhorizon.dev) answer. The PR's local end-to-end runs are in the description.
- DEFAULT_ENDPOINT is https://telemetry.playhorizon.dev. The heartbeat, the launcher's live count and the
outbox upload all go there; HORIZON_TELEMETRY_URL still names another server, and HORIZON_TELEMETRY_URL=off
sends nothing at all. CI workflows and the harnesses that start the game or the launcher set off. Consent
is unchanged: the four switches stay opt-in, only the anonymous ping is always on.
- GET /v1/benchmarks/summary: players' benchmark uploads grouped by GPU, CPU, render distance and game, groups
of 3+ installs only, medians per install then across installs, no install ID, day, OS version, driver,
memory or screen. Cached 10 minutes, CORS for the website. Same rules as website/scripts/benchmarks.mjs,
tested on the same fixtures; a deleted install leaves its groups at once.
- /v1/aggregate no longer lists the newest 50 raw benchmark reports (hardware, driver, day, one install each):
uploads are published only through the summary.
- Behind Railway's proxy the rate limit counts X-Real-IP, else the last X-Forwarded-For entry (the one the
proxy adds; a client could choose the first). Still in memory only, never stored or logged.
- crates/horizon-telemetry-server/railway.json: Dockerfile build from the repository root, health check
/healthz, one replica in europe-west4 (one SQLite file on the /data volume).
- Docs: docs/telemetry.md, the server README (the summary's shape, Railway setup), the privacy page (the ping
is on, where the data lives, how benchmark uploads are shown).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On single player's server thread the broken log's drop and its pickup
both come from the server, a few ticks after `break` answers: the
pickup could land before the scenario looked for the drop, which it
then never found (StopIteration). Step 3 now waits up to 20 ticks for
the logs in the inventory or the drop to walk to.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Step 2 clones a stack in the creative tab and closes the screen. The
clone lived only on the client: ItemPickerMenu's carried stack is the
inventoryMenu's, and AbstractContainerMenu.removed clears the carried
stack only for a ServerPlayer, so vanilla's client still holds it when
the survival inventory opens. With a stack in hand the middle button's
release sends CLONE (AbstractContainerScreen.mouseReleased); the server
finds the carried stack it was told about wrong and sends the empty one
(broadcastChanges). On single player's server thread that answer takes
a round trip, as in vanilla, so step 3 now waits up to 20 ticks for it.
The in-frame world puts the clone away when the menu closes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
From b9f3821f (fix/server-thread-default): the host window draws what
it was sent on the server thread, and agent_control_e2e waits two ticks
where the server answers a click.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On a client of a server (single player's server thread, a LAN guest) the
placed block comes with the server's answer: `place` waits up to a second
for it before reporting no change. agent_control_e2e tries the next free
spot when the tree's trunk above the broken log hides a spot's faces.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry picked from commit 9821bb7ef54112200c720c2ca91f10187b4ea693)
Low leaves can hide every face of the trunk's bottom log from where the
walk stops ("no face of block ... is visible from here"). The scenario
now breaks inside the tree loop and moves on to the next tree, as it
already did for one it cannot reach.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- survival and ui asserted that the Create New World click itself
answered `playing`. Since 7f4193f8 the in-frame world is prepared off
the UI thread behind the title screen, so the click answers `title`
and both scenarios failed in the frame (main's default) every time.
They now wait for `loaded`, as the other scenarios do.
- The survival spawn is random within the spawn radius, on the first
full top face below the MOTION_BLOCKING height
(PlayerSpawnFinder.getLevelRespawnPos), so a tree's leaves too: from
there no tree was reachable within goto's three-block drops. The
scenario now drops the player to the ground under the leaves first.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The middle button (keyPickItem) over an inventory or the creative screen no
longer clicks like the left button: with infinite materials it clones the
hovered stack at full size (CLONE) or, holding a stack, drags full stacks into
slots (quick craft type 2); otherwise it does nothing. Agents can middle-click
widgets (click ID middle), and creative slots are ui widgets now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
lab.VERSION, lab.GAME (.local/minecraft/<version>) and lab.RESOURCES
(.local/horizon-client/<version>) come from the lock, and lab.version_info()
reads the pinned client jar's version.json. Every harness, generator and
tool that named 26.3 (resource folders, jar and source paths, result
records, the vanilla client's --version and asset index, the bare login
client's protocol, the benchmark's options.txt data version, the packet id
table's header) uses them, so a candidate worktree pinned to another version
runs them unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Follow-up to #773 (consume effects on food):
- ApplyStatusEffectsConsumeEffect always draws its float and skips the
effects when the draw is >= probability (#773 skipped the draw at 1.0
and applied on <=).
- on_consume_effects run for every consumable, not only food: milk can be
drunk (clear_all_effects, then the bucket), honey removes poison and
leaves its glass bottle, stews their bowl (UseRemainder).
- The consumable's own timing, sound, animation and particles
(consume_seconds, sound, has_consume_particles) through horizon-world's
Consumable.emit, in vanilla's draw order; FoodProperties.onConsume's
sound and burp; the used statistic; the eat or drink game event after
the effects.
- teleport_randomly (chorus fruit) through horizon-world's verified
TeleportConsume port, with LivingEntity.randomTeleport for the player
(border clamp, entities_can_teleport_to ground, collision and liquid
checks), the teleport game event and sound, level event 2017's
directional particles and entity event 46's portal particles.
- tools/agent_control_e2e.py: a consume scenario (golden apple at full
hunger, milk, honey, chorus fruit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
wait-for screen NAME|widget ID|loaded|chat TEXT|event KIND replies once
the game gets there (--timeout, default 30 s, ends in an error of kind
timeout; --since SEQ counts chat and events from an earlier point), so
scripts and agents stop polling. events [--since SEQ] lists what
happened, oldest first, numbered: screens, chat lines, deaths, respawns,
items gained or lost, dimensions, worlds opened and closed. The state is
compared once per event loop turn and only under --control, so watching
never changes the game. respawn now says when the death screen's button
is not active yet instead of clicking it to no effect.
horizon-ctl and its MCP gain both; tools/agent_control_e2e.py gains an
events scenario.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
options-get [KEY...] gives every options.txt option by its vanilla key
with its value and what it accepts (type, range, step or choices; enums
by name, key bindings as key.keyboard.w). options-set KEY VALUE sets one,
applies it at once and saves options.txt as the options screen does
(language and resource packs reload too); a value the option refuses is
an error naming what it accepts, and the option keeps its value. After a
set, ui reads the redrawn screen.
horizon-ctl and its MCP gain both; tools/agent_control_e2e.py gains an
options scenario.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`ui` lists every widget the last frame drew (title menus, pause, options
in both menu styles, inventories and menus) with an id, its kind, label,
bounds and value; `click ID` clicks one through the same window events
as the mouse, `set ID VALUE` moves a slider along its track. Ids come
from labels (FOV: 70 is fov), world/server/language rows carry their
names, Horizon switches, sliders and segments their row's caption, and
inventory slots are slot-ID with their stacks. After a click, the next
widget command waits for a redrawn frame.
horizon-ctl and its MCP gain ui, click and set, and launch --offscreen.
tools/agent_control_e2e.py gains a ui scenario and runs offscreen on
macOS; both scenarios now handle Singleplayer opening Create New World
when there are no worlds.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On a virtual screen: create a world through the title menus, walk to a tree,
break a log, craft planks in the inventory screen, place one and read it
back, once with horizon-ctl commands and once over horizon-ctl mcp.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
rdbtCVS
Ludovic Marie