deepseek-harness
deepseek-ai
DeepSeek Harness: Everything is a Plugin.
PROJECT TOPICS
INSTALL REFERENCE
dsh plugin --profile web add github:ZenMux/dsh-plugins
该命令指向仓库当前默认分支;尚无绑定当前 commit 的完整验证结果。
PROJECT README
Use ZenMux models in DeepSeek Harness (DSH) without copying an API key. Sign in through your browser, search the full model catalog, and let the plugin choose the cache-capable protocol for each model.
/zenmux login, /zenmux status, and /zenmux logout commands.dsh plugin --profile web add @zenmux/dsh-plugins@latest
dsh web
In a DSH conversation, run:
/zenmux login
Approve the ZenMux authorization page and return to DSH. Check the connection with /zenmux status; disconnect this installation with /zenmux logout.
Open the model selector and search by model name or ID.
The first eligible Anthropic request can create a prompt cache. Later requests with the same prefix can report cache reads.
Quit the running DSH process, update the plugin, then restart DSH:
dsh plugin --profile web add @zenmux/dsh-plugins@latest
dsh web
Confirm the installed version with:
dsh plugin --profile web list --depth 0
An older manual model list is overriding the bundled catalog. In ~/.dsh/settings.yaml, remove only the models: arrays under llm-pi-ai.providers.zenmux and llm-pi-ai.providers.zenmux-models, then restart DSH. Keep credential references and unrelated providers.
API key is invalid or AUTHRun /zenmux login again.
OAuth requests connect directly by default and can use an explicit HTTP(S)/SOCKS proxy. Custom TLS interception may also require NODE_EXTRA_CA_CERTS before DSH starts. See the implementation and rollout notes for advanced deployment details.
pnpm install
pnpm test
Model catalogs are release snapshots generated from ZenMux's OpenAI and Anthropic model-list endpoints. Source and issue tracking live at ZenMux/dsh-plugins.
MIT License
CLASSIFICATION EVIDENCE
系统优先读取 GitHub Topics,再与站内分类词典和词根规则比对。当前命中: 无有效分类标签。