earakely-scale/RosegardenFromRequestComputerEnv-REA

Rosegarden: a computer-use environment built from a one-line request

The real Rosegarden 22.12 music sequencer running on an Ubuntu 24.04 desktop, as an agentenv computer environment. An agent completes the task only through screenshots, mouse and keyboard, and a verifier grades the saved files with no model.

Claude Code built this environment from a one-line request in about 27 minutes, with only REA's build skill and MCP tools; How it was made has the request and the steps. Rosegarden opens a small music project, Documents/sunday-jam.rg, at launch: four named tracks, Melody, Bass, Piano Chords and Drums, each with one segment, among the 32 tracks Rosegarden gives a composition. The agent has to rename the Bass track to Bass Guitar and save. It sees only the screen and acts only with the mouse and keyboard.

Watch

• Agent runs (Claude Code with Claude Sonnet 5.5): run 1, run 2, run 3
• Witness, which proves the task can be done through the computer: witness.mp4

Each video is the computer's own screen recording at 10 frames per second.

The task

In Rosegarden, rename the track called Bass to Bass Guitar in my Sunday jam project, Documents/sunday-jam.rg, and save it.

bundle/artifacts/rename-bass-track-check/verify.py reads the app's files through the computer's data/get view and passes only if every check holds:

Check	Kind	File
check-1	value_equals	Documents/sunday-jam.rg
guard-1	value_equals	Documents/sunday-jam.rg
guard-2	value_equals	Documents/sunday-jam.rg
guard-3	member_count	Documents/sunday-jam.rg

An .rg file is gzip-compressed XML, and REA reads the XML inside. The checks find tracks by their id in the composition: track 1 must be labelled "Bass Guitar", tracks 0 and 2 must keep their labels, Melody and Piano Chords, and the composition must still have 32 tracks (member_count). Every check is REA's draft, unedited: the rename, two guards on the neighbouring tracks and the track count.

Results

3 of 3 runs passed, about $0.82 per run.

Run	Outcome	Turns	Cost	Video
1	passed	12	$0.82	video
2	passed	15	$0.86	video
3	passed	13	$0.77	video

Run it

You need Docker. agentenv-framework builds the computer: Ubuntu 24.04 with the app installed from Ubuntu's archive (rosegarden 22.12.1).

pip install agentenv-framework==0.9.1431
hf download earakely-scale/RosegardenFromRequestComputerEnv-REA --repo-type dataset --local-dir rosegarden-request
agent-env run rosegarden-request/bundle --task rename-bass-track --agent claude-code --model <model>

The agent needs a model provider configured for agentenv. --task smoke only boots the computer, with no agent. bundle/README.md describes the bundle in detail.

How it was made

Claude Code built this environment from a one-line request, given only REA's build skill and REA's MCP tools:

Build me a Rosegarden environment where the agent renames a track in a small music project and saves it.

It asked only the skill's upfront questions, wrote no app-specific scripts, and finished in about 27 minutes. The steps were REA's usual ones:

1. Capture. REA launched Rosegarden on a Linux desktop and recorded its package, windows and opened files while the edit was made. It also recorded the app's state folders before and after.
2. Export. REA turned the capture into an agentenv computer_app. It installs the app, seeds the starting files, starts the app, and waits for its main window. It also serves typed views of the app's files for verifiers.
3. Task. The task's checks were validated against the capture: they fail on the starting state and pass on the captured end state.
4. Witness. In a fresh computer, the track was renamed through computer actions alone, and the checks were read at four stages. They held on the first try:

Stage	Checks
seed	fail, as expected: track 1 is labelled Bass; the guards pass
after_edit	pass
near_miss	fail, as expected: the track was renamed "Bass Guitr"
after_reset	fail, as expected: data/reset restored the seed

The agent runs came after the build: Claude Code with Claude Sonnet 5.5 (--model anthropic/claude-sonnet-5-5 --no-emulate, on an amd64 host) solved the task 3 times, each run recorded.

What's inside

bundle/      the agentenv bundle: envs/ (app module, seed files), tasks/, artifacts/ (verifier), evals/
media/       witness and agent-run recordings (MP4) with short GIF previews
results.json the runs above, with each check's result

Limits

• The checks read track labels only: Drums, the untitled tracks and the segments' contents are not checked.
• Rosegarden's floating Transport window sits over the main window and takes keyboard focus after a dialog closes, so Ctrl+S can miss; File > Save works.
• The seed holds about 250 example and instrument-library files that Rosegarden copies into ~/.local/share/rosegarden on its first run, with its templates and fonts, and an autosave copy of the project, which the capture picked up. The copied files are Rosegarden's own, under its licence, not part of this bundle's Apache-2.0 files. Rosegarden started cleanly from them in the witness and in every run.
• The computer installs whatever rosegarden Ubuntu 24.04's archive holds at build time (22.12.1 for these runs); it is not pinned.
• Three runs is a small sample: the solve rate shows the task is solvable, not how hard it is.

Licences

The bundle code, the task and the synthetic seed files are Apache-2.0. Rosegarden itself is not included: the computer installs it from Ubuntu's archive (rosegarden 22.12.1) at build time, under its own licence (GPL-2.0-or-later).