nvx

Cross-Platform Micro-VM Sandbox for Agentic Workloads

Runs with mockssource: GitHubRustMITcommit 5b37b4f159ad

Rust, MIT licensed. The project labels itself: agentic coding, agentic engineering, agentic workflows, hyperv, hypervisor, kvm, linux and microvm.

nvx runs, with stand-ins for the services it depends on. An Argusic agent installed it in 70 minutes and hit 7 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.

Dependencies installed in a venv and the release bundle downloaded/verified; the Python tooling test suite passes (454 tests), OpenVMM runs its 4562-test unit suite with only the container's NUMA-seccomp-blocked mbind tests failing, the openvmm binary answers version/help, and the kernel plus all three initramfs images and the erofs layer decode as valid media;

What the agent ended up with on a clean machine, in its own words. How this is measured

Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 86 of 100 (the mean of this project's run scores).

At a glance

verdict
Runs with mocks
Argusic Score
86 / 100
cost of the verifying run
$1.48 (measured)
recorded runs
2
last tested
stars
856
forks
67
open issues
97
watchers
7
size
119 MB
created
last push

Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.

What was measured

From the run that decided this verdict. Every line is a recorded value, not an estimate.

Time lost to each failure while testing nvx, drawn from the recorded run
Every bar is a failure the agent recorded, with the time it cost. Drawn from this run, not from anywhere else.

What went wrong, in the agent's own words

The agent's own words, worst first, never rewritten into advice.

  • container 40G disk filled during OpenVMM debug build (linker SIGBUS then ENOSPC)18 minutes
  • NVMe storage_tests panic index out of bounds len 64 when thread lands on host cpu 71/147/167 (container cpuset), sparse_mmap NUMA mbind EPERM via seccomp15 minutes
  • lxutil unit tests failed on overlayfs: 1024-byte symlink target hits ENAMETOOLONG (max 1023), and rename inode assertion 17!=3912 minutes
  • cargo xflowey restore-packages wanted root apt for git/libarchive-tools and bsdtar8 minutes
  • cargo/rustc and cargo-nextest missing for OpenVMM unit tests7 minutes
  • system pip blocked by PEP 668 externally-managed environment3 minutes
Install time
70 minutes
Cold machine to finish
78 minutes
Errors hit and fixed
7 hit, 7 fixed with no human help
How the result was proved
python -m unittest scripts.test_performance scripts.test_nvx_tools scripts.test_microvm_tests scripts.test_development_release exited 0 (454 tests, OK, 7 skipped); openvmm --version printed 'openvmm 0.2.0+gd0bdbed8f'; nvx.py run --dry-run printed the openvmm command; cargo nextest run (agent profile, --features ci) executed 4562 tests with 4560 passing, 2 failing (NUMA seccomp), 10 skipped; nvx.py download exited 0; gzip/cpio/ELF/erofs decodes of build artifacts all parsed
Model tokens used
6,822,191
Exact commit tested
5b37b4f159ad
Container image
sha256:cdd920bce783

What was not measured

The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.

Peak GPU memory
this ran on a CPU machine, so there was no GPU to sample
Behaviour over a long run
each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
Behaviour against real services
this run was verified against stand-in services only

Watch the session

The whole run, from a clean machine, at the speed it actually happened. Nothing is cut and nothing is sped up.

A recording of attempt 1, run , the attempt that decided the verdict above. The terminal recording it was made from is the evidence, stored with a sha256 fingerprint; the video is a rendering of it. Nothing loads from the video host until you press play.

The end of the recording

The last lines of the terminal session that decided this verdict, exactly as they were recorded.

    {"msg": "microVM guest boot cannot run: no /dev/kvm or /dev/mshv in container and this OpenVMM …
  ],
  "test_depth": "tier2_mock",
  "notes": "No third-party keys required. Python unit suite fully green (454 tests, 7 Windows-only …
  "verified_how": "python -m unittest scripts.test_performance scripts.test_nvx_tools scripts.test_…
  \"final_state\": \"Dependencies installed in a venv and the release bundle downloaded/verified; t…
}
print(\"ARGUSIC_RESULT:\" + json.dumps(report))
EOF" in /work/repo
 succeeded in 0ms:
tokens used
6,822,191

Replay the whole session, every command from a clean machine to this point.

How it was tested

One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.

Strengths and limits

Measured facts, not opinions. How this is written.

What went well

  • Recovered from all 7 errors without a human stepping in, which says the failures are documented well enough to solve.
  • Licensed MIT, as reported by its host.
  • Measured 2 times, so the result is not a one-off.

What did not, or is not known

  • Only reached a running state against stand-in services, so nothing here proves it works against the real ones.
  • Took 70 minutes to install, slower than the median of the 29 comparable projects Argusic has measured.

Of the 30 agentic-coding projects Argusic has installed and timed, nvx was the 30th fastest to reach a running state, and 23 of 30 reached one at all.

Also tested, in the same area

Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.

Run history

tested byattemptoutcomescorecost USDwhen
Argusic Runner2/3Unverified80.000.05
Argusic Runner1/3Runs with mocks92.001.48

Topics (from GitHub)

agentic-codingagentic-engineeringagentic-workflowshypervhypervisorkvmlinuxmicrovmsandboxvm

Embed the badge

Markdown for the project README. It links back here; terms on the terms page.

[![Tested by Argusic](https://argusic.com/badge/nvx.svg)](https://argusic.com/subject/nvx)

Questions

Does nvx run?
nvx runs, with mocks standing in for external services. Argusic installed and launched it on a clean machine in 70 minutes, hitting 7 errors on the way, and recorded the session.
How did Argusic test nvx?
On a fresh, disposable machine, with every command recorded and the repository pinned at commit 5b37b4f159ad. 2 attempts are recorded, and the full method is on the methodology page.
What did it cost to test nvx?
The run that produced this verdict cost $1.48: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
How long does nvx take to install?
70 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
Does nvx need a GPU?
Not to start. Argusic reached a running state on a machine with no GPU.
What goes wrong when installing nvx?
7 things broke in the recorded run, and 7 were fixed without human help. Each one, and the time it cost, is listed on this page.
How does nvx compare with the alternatives?
Of the 30 agentic-coding projects Argusic has installed and timed, nvx was the 30th fastest to reach a running state, and 23 of 30 reached one at all.
Where is the evidence for nvx?
All 2 recorded runs are on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.

Discussion