DeepSeek Harness 的原生 Rapid-MLX 提供者：dsh 自動發現已部署模型並在工作階段內管理，上下文壓縮貼合 Mac 統一記憶體上限。
DSH 整合
原生執行環境
作者聲明
安全稽核
未稽核
最後核實
2026-08-30
授權條款
Apache-2.0
01它能幫你完成什麼？
Run local Rapid-MLX models inside DeepSeek Harness without hand-maintaining per-model facts
`dsh` auto-discovers served models from `/v1/models` — context window, reasoning/tool parsers, MoE/hybrid, modalities — with no hand-written settings.yaml
DeepSeek Harness users on Apple Silicon running Rapid-MLX who want accurate, server-driven model facts
See, pull, remove and health-check Rapid-MLX models without leaving the dsh session
Five agent tools — `rapid_mlx_serving`, `rapid_mlx_cached`, `rapid_mlx_pull`, `rapid_mlx_remove`, `rapid_mlx_health` — plus a `/rapid-mlx` command
Agents and developers who want to manage served/downloadable models and check health in-session
Get truthful reasoning controls and machine-fitted context compaction
Reasoning selector only appears for models that actually have a reasoning parser; compaction timed to the server's `max_model_len` (unified-memory ceiling), not a hand-written number
Developers hit by silent failures from stale context windows or dead reasoning selectors when switching Rapid-MLX models
02如何將外掛接入 DeepSeek Harness？
先決條件
- Node ≥ 22.15 (dsh imports Node's Zstd stream API without declaring it)
- a running Rapid-MLX server
安裝步驟
- 01
$ dsh plugin --profile web add @raullenchai/dsh-provider
- 02
$ dsh plugin --profile web add github:raullenchai/rapid-mlx-dsh-provider
- 03
export RAPID_MLX_BASE_URL=http://localhost:8000/v1 # optional; this is the default
- 04
$ dsh web
- 05
# $DSH_HOME/settings.yaml agent-default-model: provider: rapid-mlx model: qwen3.6-35b-8bit
驗證整合成功
- Installs and activates as a profile layer; the entry shows up in `dsh --profile headless --dump-config` with no "declares no dsh.bundle" warning.
- Registers the `rapid-mlx` route with `ctx.llm` and serves real queries.
03DSH 整合程度與能力邊界
Native dsh LLM adapter: installed via `dsh plugin --profile web add`, registers the `rapid-mlx` route and reads model facts from the Rapid-MLX server's `/v1/models`.
Server-driven model discovery
Rapid-MLX `/v1/models` HTTP endpoint→served models with context window, reasoning/tool parsers, MoE/hybrid, modalities — deduped
In-session model management tools
the active dsh agent session→five tools (`rapid_mlx_serving`, `rapid_mlx_cached`, `rapid_mlx_pull`, `rapid_mlx_remove`, `rapid_mlx_health`) plus a `/rapid-mlx` command
`rapid_mlx_pull` and `rapid_mlx_remove` change on-disk cached models (non-interactive, forced `-y`)Memory-fitted context compaction
dsh-compaction-basic compaction request→compacts at `thresholdRatio × capacity` (0.8 default) using server `max_model_len` when available, else `context_window`
Conformant LLM adapter (cookbook contract)
dsh LLM calls — chat, tool calls, streaming, abort→streaming responses meeting the official adapter protocol obligations
04適合誰？何時不該用？
適合
- DeepSeek Harness users on Apple Silicon running Rapid-MLX who want accurate, server-driven model facts
- Agents and developers who want to manage served/downloadable models and check health in-session
- Developers hit by silent failures from stale context windows or dead reasoning selectors when switching Rapid-MLX models
不適合
- The route is registered as `rapid-mlx`. If your settings.yaml also declares a `rapid-mlx` provider under `llm-pi-ai`, the two compete for one route name (`registerAdapter` owns provider exclusivity). Use one or rename ours.
- Several server-reported facts are read but not yet acted on — `recommended_sampling`, `tool_call_parser` (no fast-fail on models that can't emit tool_calls), and `is_hybrid`/`is_moe`/`capabilities`; images are refused with `UNSUPPORTED` rather than carried.
05相容性、維護與安全提醒
- The route is registered as `rapid-mlx`. If your settings.yaml also declares a `rapid-mlx` provider under `llm-pi-ai`, the two compete for one route name (`registerAdapter` owns provider exclusivity). Use one or rename ours.
- Several server-reported facts are read but not yet acted on — `recommended_sampling`, `tool_call_parser` (no fast-fail on models that can't emit tool_calls), and `is_hybrid`/`is_moe`/`capabilities`; images are refused with `UNSUPPORTED` rather than carried.
- DSH is still a developer preview that moves fast; dsh 0.1.0-rc.8 is API-compatible with rc.7 but the author treats compatibility as tracking a moving target, not a frozen promise.
Apache-2.0 · published to npm as @raullenchai/dsh-provider (latest release v0.2.0, 2026-08-19)
06常見問題
如何為 DeepSeek Harness 安裝 dsh-provider？
執行 `dsh plugin --profile web add @raullenchai/dsh-provider`（或 `dsh plugin --profile web add github:raullenchai/rapid-mlx-dsh-provider` 裝來源）。需要 Node ≥ 22.15 與一個正在執行的 Rapid-MLX 伺服端，接著執行 `dsh web`。
安裝前提為何？
Node ≥ 22.15（dsh 會直接 import Node 的 Zstd 串流 API 而未宣告）以及一個正在執行的 Rapid-MLX 伺服端。僅當伺服端不在預設位址 http://localhost:8000/v1 時才需設定 `RAPID_MLX_BASE_URL`。
它如何接入 DeepSeek Harness？
它以原生 LLM 配接器身分註冊為 `rapid-mlx` 路由；dsh 連線本地 Rapid-MLX 伺服端的 OpenAI 相容 `/v1` 端點，並從 `/v1/models` 讀取模型資訊。在 settings.yaml 中將 `agent-default-model` 指向 `provider: rapid-mlx` 即可。
與 dsh 內建 openai-completions 路由有何不同？
通用路由要你手寫每個模型的 facts 到 settings.yaml，且只懂你填的內容。本配接器讀取 `/v1/models`，切換模型不需重新設定，且推理檔位只在實際具備推理解析器的模型上出現。
排障：我的 settings.yaml 已有 rapid-mlx 提供者該怎麼辦？
兩者都會註冊 `rapid-mlx` 路由並爭搶同一路由名（`registerAdapter` 擁有提供者排他性）。只用其中一個，或重新命名本插件的路由。注意 dsh 0.1.0-rc.8 與 rc.7 API 相容，但 DSH 仍是快速迭代的開發者預覽版。
07相關的 DSH 工作流程
dsh-routing-suite
作者 yjh051108
DeepSeek Harness（dsh）路由套件：先裝執行時注入器，再載入任務感知的推理模式路由預設（已實測 P1-P23）。
brooks-lint
作者 hyhmrright
基於 12 本經典工程書籍的 AI 程式碼評審工具，提供腐化風險診斷、書籍引用、嚴重度標籤及 6 種分析模式。
dsh-plugin-shop
作者 livxue
DeepSeek Harness 最全面的外掛市場，每日更新、全網採集、釋出前稽核。
dsh-our-free-model
作者 zouyuxuan122
在 dsh 裡裝上這個外掛即可，無需登入、註冊或填 API Key，就能使用包括 Muse Spark 1.3、MiMo V2.6 在內的前沿模型——完全免費，不限量。 All you do is install this plugin in dsh: no login, no sign-up, no API key — the frontier models are just there, Muse Spark 1.3 and MiMo V2.6 among them. Completely free, with no usage cap.
08資料與來源
A native [Rapid-MLX](https://github.com/raullenchai/Rapid-MLX) provider for
Registers the `rapid-mlx` route with `ctx.llm` and serves real queries.
此頁面根據專案公開文件、儲存庫中繼資料與 DSH Plugins 的結構化解析所產生；最後核實於 2026-08-30。發現錯誤？提交更正。
最佳 DeepSeek Harness 外掛
從全目錄挑出的 12 個值得優先安裝的外掛，涵蓋各個分類。