dsh-model-search-select

1. 安装并启动 DeepSeek Harness：npx @deepseek-ai/dsh web
2. 在终端执行上面的安装命令（CLI 会解析插件并核验来源）
3. 用 dsh plugins list 确认已安装，必要时重启 Harness 生效

插件以当前 dsh 进程的权限运行，安装时可能执行代码。请先通读仓库源码与许可证，确认无破坏性命令与越权访问；本站只做索引，不对第三方插件安全性作担保。

事实信息来自公开插件目录快照（2026-10-01），介绍文案由本站再加工。

以下为插件仓库 README 全文（原始内容，由公开目录抓取整理）。

# dsh-model-search-select

> Community plugin for [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness) — **not an official DeepSeek product.** No endorsement, review, or store listing by DeepSeek.

Searchable model selector for the DSH web UI.

A faithful fork of the native model seat: same two-pane **Model / Effort** menu,
provider-grouped model list, and reasoning-effort selection — **plus a search box**
at the top of the model list that filters while you type (matches model name, id,
description, and provider name).

The shipped selector is not removed: this plugin registers on the same
`conversation.input.model` seat with a lower priority, so it wins the seat while
active, and the native selector stays mounted underneath as a fallback.

## Install

```sh
dsh plugin --profile  add dsh-model-search-select
```

Then restart the profile:

```sh
dsh --profile  web
```

Open the model selector in the composer (bottom-right). In the **Model** pane a
search box appears at the top; type to filter. The **Effort** pane is unchanged.

## Uninstall

```sh
dsh plugin --profile  remove dsh-model-search-select
```

Restart the profile: the native selector returns (it was never removed).

## Requirements

- DSH (any profile with the web UI)
- The model directory service (`modelDirectories`), shipped by the web bundle.

## How it works

- `lib/index.js` — host half (no-op; the seat is pure client UI).
- `lib/client.js` — client bundle. Registers on `conversation.input.model`
  with `priority: -1`, reading the same shared `ModelDirectory` the native seat
  uses, so a picked model lands on the standard `directory.select()` path
  (provider + model + reasoning effort).

## License

MIT