DeepSeek Harness (dsh): Version, Updates, Stability (2026)

dsh is still on 0.1.0-rc.6, the build it launched with, and the public repository has not received a commit since 2026-08-13. That is four days of no visible upstream change against a README that describes the project as iterating rapidly.

None of that makes it a bad tool. It does change what you should do before wiring it into anything you depend on.

Current version:    0.1.0-rc.6, published 2026-08-13 12:35 UTC
Versions published: 6, all between 2026-08-10 and 2026-08-13
Last public commit: 2026-08-13 11:38 UTC
Releases / tags:    0 / 0
Branches:           master only
Issue tracker:      disabled, Discussions and Discord only
Discussions:        2,713 threads (snapshot 2026-08-17)
License:            MIT
Official status:    developer preview, breaking changes promised
Entry point:        npx @deepseek-ai/dsh web, Web UI on 127.0.0.1:3080

What Version of DeepSeek Harness Is Current?

0.1.0-rc.6, published 2026-08-13 at 12:35 UTC. On npm both the latest and next dist-tags resolve to it, so a plain npx @deepseek-ai/dsh and an explicit @next give you the same bytes.

The full published history is six versions across four days, with half of them landing on launch day itself:

Every release predates or matches the public launch. The version you install today is the version that shipped on day one.

Has DeepSeek Harness Been Updated Since Launch?

Not in public. The last commit on master is dated 2026-08-13 11:38 UTC and the repository’s own pushed_at timestamp is 2026-08-13 13:00 UTC. Checked again on 2026-08-17, both are unchanged.

There is also nothing else to read state from:

• 0 releases and 0 tags, so there is no upstream signal for “this build is good”
• master is the only branch, so no visible development line
• npm has not moved either, which rules out “they ship to npm and backfill git later”

This matters more than it would for a mature project, because the README promises breaking changes. A tool that will break you, and gives you no tag to pin against, leaves the npm version string as your only anchor.

How Do You Check dsh’s Status Yourself?

Three commands, no GitHub account needed. Every number in this article moves, so verify rather than trust a snapshot.

Current version and the full publish history:

npm view @deepseek-ai/dsh dist-tags versions time

Whether the public repository has moved at all:

curl -s https://api.github.com/repos/deepseek-ai/deepseek-harness \
  | jq '{pushed_at, has_issues, open_issues_count}'

That returns 2026-08-13T13:00:21Z, false and 0 as of 2026-08-17. Releases and tags are separate calls and both come back as empty arrays:

curl -s https://api.github.com/repos/deepseek-ai/deepseek-harness/releases
curl -s https://api.github.com/repos/deepseek-ai/deepseek-harness/tags

If pushed_at has moved past 2026-08-13 by the time you read this, the picture below has changed and pinning your version matters more, not less.

Is DeepSeek Harness Production Ready?

No, and DeepSeek says so directly. The README carries its own heading for it:

DeepSeek Harness is currently in developer preview and is iterating rapidly. THERE WILL BE COMPATIBILITY-BREAKING CHANGES.

Two structural facts back that up rather than contradict it.

Issues are disabled on the repository, so the only official channels are Discussions and Discord. Bug reports and feature requests land in the same feed with no assignee, no label and no status. That is a fine setup for gathering signal and a poor one for tracking a regression you filed.

The plugin architecture points the same way. Everything in dsh is a Cordis plugin, including the UI, which is what makes the ecosystem move fast. It also means a breaking change upstream can land in a surface you did not know you depended on.

Why Does the dsh Repo Say “Iterating Rapidly” With No Commits?

Because what got published is the code, not the process around it. The repository was created on 2026-08-13 but arrived carrying its full history: 12,293 commits reaching back to 2026-06-10, whose first commit reads “Initialize repo with README, AGENTS.md, and CLAUDE.md symlink.” Two months of work, all of it visible.

What stayed private is everything around that code. The merge commits reference pull requests #2519, #2520 and #2521 in a separate deepseek-harness organization you cannot open, there are no branches besides master, and nothing has been pushed since launch day.

So the code history is unusually transparent and the current state is unusually opaque. You can read every line of how dsh got to 0.1.0-rc.6, and nothing at all about what happens next.

“Iterating rapidly” is probably true, then, and simply not observable from outside. That is a legitimate way to run a project.

It has one concrete consequence for you. For anything after 2026-08-13, you cannot tell “quiet because stable” apart from “quiet because busy.” You will find out at the same moment everyone else does, when the next version appears on npm.

The practical response is to pin. npx @deepseek-ai/dsh@0.1.0-rc.6 costs nothing today and is the only thing standing between you and a promised breaking change.

How Big Is the Community Around dsh?

Large and growing much faster than the code. GitHub Discussions held 2,713 threads when we checked at 08:00 UTC on 2026-08-17, against about 620 in our own snapshot three days earlier. It passed 2,730 within the hour we were writing this. Star and fork counts move just as fast, so treat any specific number here as a timestamp, not a fact; the direction is the only stable part.

What the volume is made of:

The gap between 2,713 discussion threads and zero commits is the most useful number pair here. Demand is real. Visible iteration is not.

Note what the ecosystem column implies. When plugins ship faster than the core they plug into, and the core promises breaking changes, the plugins are the first thing that breaks. That is an argument for keeping your dsh install thin rather than for avoiding it.

Does dsh Show You What Each Task Costs?

Not today. There is no per-turn token or cost display in the shipped build. One of the more upvoted Ideas threads, filed 2026-08-14, is a request to add exactly that, and a community cost-tracker plugin appeared on launch day to fill the gap.

The timing makes this sharper than it looks. DeepSeek moved its whole V4 family to peak and off-peak pricing on 2026-08-16, so the same agent loop now costs one of two prices depending on the hour.

Cache hits on deepseek/deepseek-v4-pro went up 12.1x at peak, and deepseek/deepseek-v4-flash 5x. An agent harness that resends the same system prompt and file contents every turn is precisely the workload that change targets.

Until the harness surfaces it, your options are the provider dashboard, your own proxy logs, or one of the community plugins built to show cost per task and peak-hour status. None of those ships with dsh.

When Is dsh Safe to Use Today?

Safe to explore, not safe to depend on. The split is cleaner than the developer-preview label suggests, because it comes down to what breaks when the next version lands.

The dividing line is whether an upgrade that breaks your setup costs you an afternoon or a deadline. DeepSeek has told you in capital letters which of those it is planning.

What Should You Do Before Wiring dsh Into Your Workflow?

Pin the version, keep the model endpoint swappable, and do not put it on the critical path yet. The first two cost nothing. The third is the actual decision.

A concrete checklist:

1. Pin it. Run npx @deepseek-ai/dsh@0.1.0-rc.6, never bare dsh.
2. Point it at a base URL you control, not a hardcoded vendor default.
3. Keep a second harness working for anything with a deadline.
4. Watch npm rather than GitHub, since npm is where the next version appears first.
5. Reread the README after any upgrade. Breaking changes are announced there and nowhere else.

Point 2 is the one people skip. dsh reads DEEPSEEK_BASE_URL and DEEPSEEK_API_KEY for its built-in route, and the custom-provider form accepts any OpenAI-compatible endpoint, so the model layer is genuinely swappable if you set it up that way from the start.

On ofox that means deepseek/deepseek-v4-flash and deepseek/deepseek-v4-pro behind one key. The value is not the gateway itself. It is that on the day dsh breaks, or the day pricing moves again, the model string is the only thing you edit.

The step-by-step for both routes is in the dsh custom provider setup guide. For how dsh sits against Claude Code, Codex and the rest on extensibility and model lock-in, see the coding agent harness comparison.

References

• deepseek-ai/deepseek-harness on GitHub
• dsh README, developer preview notice
• DeepSeek Harness Discussions
• @deepseek-ai/dsh on the npm registry
• Cordis, the plugin runtime dsh is built on
• DeepSeek API pricing
• DeepSeek change log