@dsh-workbench/plugin
Worka — notes, tools, and ongoing agent work for DeepSeek Harness.
AI 分析
适合需要持久化工作台状态、构建交互式笔记的重度AI用户。
インストール
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:kllilizxc/dsh-workaドキュメント
README 全文を読む ↗Configuration and storage
The bundle registers the Worka plugin and a SQLite backend in the host. Defaults:
| Setting | Default | Purpose |
|---|---|---|
dataRoot | workbench under DSH home | Note files, builds, and runtime metadata |
SQLite path | workbench/state.sqlite under DSH home | Persistent Worka state |
workspaceRoot | Host working directory | Default workspace for new conversations |
bunPath | bun | Executable used to build notes |
desktopNotifications | false | Optional macOS notifications |
tickIntervalMs | 1000 | Scheduler polling interval |
Override the bundle in your profile's cordis.patch.yml, for example:
- id: workbench
config:
workspaceRoot: /absolute/path/to/projects
bunPath: /absolute/path/to/bun
desktopNotifications: true
When changing the storage location, update both workbench.dataRoot and workbench-sqlite.path. Keep one owner for each Worka database; do not point two running hosts at the same data directory.
Models and API keys are configured in DSH. Worka does not create a separate credential file or default model. Switching a default model affects new conversations; existing conversations use their native DSH model selector.