earakely-scale/GnumericFromRequestComputerEnv-REA

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

The real Gnumeric 1.12 spreadsheet 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 20 minutes, with only REA's build skill and MCP tools; How it was made has the request and the steps. Gnumeric opens a small household budget, Documents/household-budget.xml, at launch: rent, groceries, utilities and transport by month from January to April, with a Total row. The agent has to change March's Groceries figure from 405.10 to 437.80 and save, leaving every other figure alone. 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 my household budget spreadsheet, Documents/household-budget.xml, change the Groceries figure for March to 437.80 and save it.

bundle/artifacts/update-march-groceries-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
march-groceries	value_equals	Documents/household-budget.xml
february-groceries-untouched	value_equals	Documents/household-budget.xml
april-groceries-untouched	value_equals	Documents/household-budget.xml
march-utilities-untouched	value_equals	Documents/household-budget.xml

The seed is saved as Gnumeric XML uncompressed (.xml) rather than Gnumeric's default gzip-compressed .gnumeric, so REA reads it as XML. Each check finds a cell by its Row and Col attributes, never by position, and compares its value within 0.001, because Gnumeric writes numbers with 17 significant digits (437.80000000000001). March Groceries must be 437.8, and February and April Groceries and March Utilities must keep their values, so an edit to a neighbouring cell fails. REA's draft suggested checks on where the cursor was left rather than on the edited cell; Claude Code wrote these instead.

Results

3 of 3 runs passed, about $0.52 per run.

Run	Outcome	Turns	Cost	Video
1	passed	10	$0.43	video
2	passed	10	$0.69	video
3	passed	10	$0.45	video

Run it

You need Docker. agentenv-framework builds the computer: Ubuntu 24.04 with the app installed from Ubuntu's archive (gnumeric 1.12.56).

pip install agentenv-framework==0.9.1431
hf download earakely-scale/GnumericFromRequestComputerEnv-REA --repo-type dataset --local-dir gnumeric-request
agent-env run gnumeric-request/bundle --task update-march-groceries --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 Gnumeric environment where the agent updates one month's figure in a small household budget spreadsheet and saves it.

It asked only the skill's upfront questions, wrote no app-specific scripts, and finished in about 20 minutes. The steps were REA's usual ones:

1. Capture. REA launched Gnumeric 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 edit was made through computer actions alone, and the checks were read at four stages. They held on the first try:

Stage	Checks
seed	fail, as expected: March Groceries is 405.10
after_edit	pass
near_miss	fail, as expected: March Groceries was typed as 473.80
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 Documents/household-budget.xml as Gnumeric XML uncompressed. A correct edit saved under another name or format, such as Gnumeric's default .gnumeric, fails.
• The Total row is not checked: Gnumeric stores its =SUM formulas, not their results.
• The computer installs whatever gnumeric Ubuntu 24.04's archive holds at build time (1.12.56 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. Gnumeric itself is not included: the computer installs it from Ubuntu's archive (gnumeric 1.12.56) at build time, under its own licence (GPL-2.0-only or GPL-3.0-only).