LLM WORLD

DeepSeek Harness developer preview

DeepSeek · · Coding harness

DeepSeek open-sourced its first first-party agent harness under MIT as a v0.1 developer preview. Built on a dependency-injection core called Cordis, every layer — model adapter, tool registry, session log, sandbox, web UI and the agent loop itself — is a replaceable plugin, and it ships adapters for Anthropic, OpenAI, Bedrock, Azure and Google rather than only DeepSeek's own models.

Why it matters

The last of the large Chinese labs to ship a harness, and the one that refused to make it a product: released as a preview, provider-agnostic by design, with DeepSeek's own warning not to build anything load-bearing on it yet. It passed 160,000 GitHub stars within days, making the harness — not the model — DeepSeek's most-starred release.

This is LLM World's historical reading, not a claim by DeepSeek. The facts above and the sources below are what it rests on.

Sources

Record status: partly verified · confidence 70/100

Two independent outlets published same-day coverage of the release on 13 August 2026 — The New Stack (dateline 'August 13th, 2026 at 1:20pm') and CryptoBriefing (byline 'Aug. 13, 2026') — and the deepseek-ai/deepseek-harness repository was created 2026-08-13T11:56:32Z, which corroborates but is not the basis for the date. DeepSeek's own product page at deepseek.com/harness carries no date at all, only a 2026 copyright, so there is no primary dateline. Note the only tagged release, dsh-v0.1.0-rc.7, was published 2026-08-17 — four days after the announcement — and was deliberately not used as the date. Also note this launched alongside DeepSeek-V4-Pro's API GA, which is a model event belonging in the release dataset, not here.