dsh-kimi
Run the following command in DeepSeek Harness:
dsh plugin install myk5010/dsh-kimi
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install myk5010/dsh-kimi in the DeepSeek Harness terminal to install this plugin; the source repository is at https://github.com/myk5010/dsh-kimi
About this plugin
Using Kimi models inside DSH used to mean hand-editing provider configs, stitching together an OAuth flow, and keeping an eye on token expiry - each step a chance for something to go wrong. dsh-kimi bundles all of that into a single plugin: install it and Kimi shows up as a first-class provider in the model picker, no manual configuration required.
Under the hood it registers a provider route over the OpenAI Responses protocol with four models (kimi-for-coding, kimi-for-coding-highspeed, k3 with 1M context, and k3-256k). Login runs through a device-code OAuth flow that dsh owns independently, so token refresh here never evicts your Kimi Code CLI session. An every-5-minute timer keeps the access token alive automatically, and three agent tools (kimi_login_begin, kimi_login_status, kimi_login_refresh) let you manage the whole lifecycle from a conversation.
Built for developers who want Kimi coding models available in any DSH profile - web, headless, or otherwise - without manually maintaining provider configs or tracking token lifetimes.
Use Cases
- Pick Kimi coding models directly in the DSH model picker without hand-editing provider config
- Log in, check status, and refresh tokens via agent tools in conversation - no terminal switching
- Call the same set of Kimi models consistently across web and headless DSH profiles
Best For
- Developers using DSH who want to add Kimi coding models without manual setup
- Engineers who prefer not to manage OAuth token lifetimes and expiry refresh by hand
- Teams that need a consistent model integration across multiple DSH profiles
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.