Back to directory

kimix

Maintenance: Active

sikao-engine/kimix

The next-gen lightweight coding agent cli

View on GitHub
$ dsh plugin add kimix

111

stars

12

forks

Python

Language

2026-03-13

Created

2026-08-25

Last push

A Python coding-agent CLI that uses DeepSeek (and 25+ other LLM backends) via `kimix --config=xx.json`; install from source with `python install.py` or via `pip install kimix`.

DSH integration

Compatible

Author-claimed

Safety audit

Unaudited

Last verified

2026-08-25

License

Not specified by the author

01What can it help you accomplish?

  • Run an LLM-backed coding agent that drives Shell, Glob and other tools to complete real development tasks

    Code changes and validated config files, with subprocess / sub-agent / background-task execution and format-check warnings

    Developers who want a lightweight coding-agent CLI that can be pointed at DeepSeek (or OpenAI / Anthropic / …) models

  • Switch between DeepSeek and 25+ other LLM providers without changing your workflow

    One KimiX CLI that routes prompts to any verified backend (deepseek, openrouter, openai, anthropic, gemini, …) through a custom config

    Teams standardized on DeepSeek that also want OpenAI / Anthropic-compatible fallback providers in a single tool

02How to install into DeepSeek Harness

Installation steps

  1. 01

    From source: run `python install.py`

  2. 02

    Via pip: run `pip install kimix`, then `python -m kimix.cli` (or just `kimix`)

  3. 03

    After setup, point KimiX at a backend config: `kimix --config=xx.json` (default templates live in `docs/`)

Verify the integration

Not specified by the author

03DSH integration and capability boundaries

DSH integrationCompatible

Runs as a standalone coding-agent CLI and points KimiX at DeepSeek (and 25+ other) LLM backends via a config file and API key — `kimix --config=xx.json`.

  • Lean system prompts

    initial prompts and tool descriptionscompressed prompts (~2000 tokens) that still cover nearly all built-in tools

  • Hardened permissions & validation

    Shell, Glob and other tool callsvalidated tool execution that reduces retry loops from failures

  • Multi-API & verified backends

    a custom config (`kimix --config=xx.json`) compatible with OpenAI / Anthropicrouting to DeepSeek and 25+ verified providers (kimi, openai, anthropic, gemini, …)

04Who is it for? When not to use it?

Good for

  • Developers who want a lightweight coding-agent CLI that can be pointed at DeepSeek (or OpenAI / Anthropic / …) models
  • Teams standardized on DeepSeek that also want OpenAI / Anthropic-compatible fallback providers in a single tool

05Compatibility, maintenance and safety notes

2026-03-132026-08-24dsh-plugin

Python · actively maintained (latest release dsh-plugin, 2026-08-14)

06Frequently asked questions

How does KimiX connect to DeepSeek Harness?

The README presents KimiX as a standalone coding-agent CLI and lists DeepSeek among its verified LLM backends; it does not describe a native DeepSeek Harness runtime integration. You connect it to DeepSeek by selecting DeepSeek in a config file (`kimix --config=xx.json`).

Do I need my own API key?

Yes. The README states the repo supports various API keys (OpenAI, Anthropic, and others) and uses `kimix --config=xx.json` after setup, so you bring your own DeepSeek (or other provider) key.

Can I use models other than DeepSeek?

Yes. KimiX is tested against kimi, xai, anthropic, openai_legacy, openai_responses, google_genai/gemini, vertexai/vertex and 25+ Hermes-ported providers including deepseek, openrouter, xiaomi, zai, minimax and bedrock.

How do I install KimiX?

From source run `python install.py`, or via pip run `pip install kimix` then `python -m kimix.cli` (or just `kimix`). The README does not list a separate verify or rollback step.

What makes it different from the original Kimi-CLI?

It compresses system prompts to ~2000 tokens, hardens Shell / Glob validation to cut retry loops, redirects large subprocess output to temp files, and adds programmable prompts plus explicit multi-task conversation management.

08Data and sources

  • Author-claimedgithub.com0e598f5aab4f…

    Tested against kimi, xai, anthropic, openai_legacy, openai_responses, google_genai/gemini, vertexai/vertex, plus 25+ Her…

  • Author-claimedgithub.com0e598f5aab4f…

    This repo supports not only KIMI LLM but also various API keys! Like OpenAI, Anthropic, etc. Default config templates ar…

This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-25. Found an error? Submit a correction.

🏆

Best DeepSeek Harness Plugins

Twelve plugins worth installing first — picked from the whole catalog, across every category.

DSH Plugins is an independent community directory of DeepSeek Harness plugins. Not affiliated with or endorsed by DeepSeek. Third-party plugins are not security-audited — review the source before installing.

New DeepSeek Harness plugins, weekly. No spam.