deepseek-harness
DeepSeek Harness: Everything is a Plugin.
Runssource: GitHubhomepageTypeScriptMITcommit 0a53fb55bea1
TypeScript, MIT licensed. The project labels itself: ai agents, cordis, dsh and dsh plugin.
deepseek-harness runs. An Argusic agent installed it in 3 minutes and hit 3 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
pnpm install and build succeed; all 17267 unit tests pass (1050 files); the dsh CLI prints help; the web UI server starts and serves a 24KB HTML page on HTTP 200 with auth token. 1 IPv6 test and 115 other tests are skipped (container sandbox limitations).
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 80 of 100 (the mean of this project's run scores).
At a glance
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.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- Container umask 0002 causes mkdirSync with { recursive: true } (no explicit mode) to create group-writable (0775) directories. The sweep's isTrustedDirectory check requires mode & 0o22 === 0, rejecting them. Affected: 9 tests in spill-local and 1 in loader-composition.10 minutes
- Boundary test (keeps a file exactly at the cutoff) is flaky because utimesSync stores second-level precision but Date.now() has millisecond precision. The sweep comparison mtimeMs >= cutoffMs can fail when sub-millisecond rounding changes the stored timestamp.5 minutes
- IPv6 loopback (::1) is unavailable in this container, causing listen EADDRNOTAVAIL on the IPv6 listener test.2 minutes
- Install time
- 3 minutes
- Cold machine to finish
- 48 minutes
- Errors hit and fixed
- 3 hit, 3 fixed with no human help
- How the result was proved
- pnpm run test exited 0, 1050 passed, 17267 tests passed, 9 skipped, 116 skipped. dsh CLI --help printed usage. 'dsh web --no-open' started server; curl with token returned HTTP 200 (303 redirect -> 200), content-type text/html, 24KB HTML body.
- Model tokens used
- 496,122
- Exact commit tested
- 0a53fb55bea1
- 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
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 verdict above comes from a different attempt, so read this as one session and not as the current result. 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.
Let me be careful about the test_depth. The tests ran with real behavior, no mocks were set up for … Wait, but the e2e tests (that need DEEPSEEK_API_KEY) weren't run. The `pnpm run test` command runs … Actually let me reconsider - the "test" command `pnpm run test` runs `vitest run` which uses the vi… For the launch: I verified the CLI works (prints help) AND the web UI works (serves HTML on HTTP 20… 2026-09-20T11:04:02.686509Z ERROR codex_core::session: failed to record rollout items: thread 01a0b… tokens used 496,122 Actually wait - I need to be more careful about the time. The user asks for install_minutes - that'… Let me be careful about the test_depth. The tests ran with real behavior, no mocks were set up for … Wait, but the e2e tests (that need DEEPSEEK_API_KEY) weren't run. The `pnpm run test` command runs … Actually let me reconsider - the "test" command `pnpm run test` runs `vitest run` which uses the vi… For the launch: I verified the CLI works (prints help) AND the web UI works (serves HTML on HTTP 20…
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
- Reached a running state on a clean machine, with the session recorded.
- Installed in 3 minutes, faster than the median of the 11 comparable projects Argusic has measured.
- Recovered from all 3 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
- Measured 6 times, so the result is not a one-off.
Of the 12 ai-agents projects Argusic has installed and timed, deepseek-harness was the 4th fastest to reach a running state, and 8 of 12 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
- A quick evaluation: it was running 3 minutes after the clone.
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 by | attempt | outcome | score | cost USD | when |
|---|---|---|---|---|---|
| Argusic Runner | 3/3 | Runs | 100.00 | 0.16 | |
| Argusic Runner | 2/3 | Ran out of time | 0.00 | 0.27 | |
| Argusic Runner | 1/3 | Runs | 100.00 | 0.20 | |
| Argusic Runner | 1/3 | Runs with mocks | 88.00 | 6.50 | |
| Argusic Runner | 2/3 | Did not run | 20.00 | 0.14 | |
| Argusic Runner | 1/3 | Runs with mocks | 92.00 | 0.57 |
Topics (from GitHub)
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/deepseek-harness)Questions
- Does deepseek-harness run?
- Yes. deepseek-harness runs. Argusic installed and launched it on a clean machine in 3 minutes, hitting 3 errors on the way, and recorded the session.
- How did Argusic test deepseek-harness?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 0a53fb55bea1. 6 attempts are recorded, and the full method is on the methodology page.
- What did it cost to test deepseek-harness?
- The run that produced this verdict cost $0.16: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does deepseek-harness take to install?
- 3 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 deepseek-harness need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing deepseek-harness?
- 3 things broke in the recorded run, and 3 were fixed without human help. Each one, and the time it cost, is listed on this page.
- How does deepseek-harness compare with the alternatives?
- Of the 12 ai-agents projects Argusic has installed and timed, deepseek-harness was the 4th fastest to reach a running state, and 8 of 12 reached one at all.
- Where is the evidence for deepseek-harness?
- All 6 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.