dsh-workspace-digest
Run the following command in DeepSeek Harness:
dsh plugin install cui-stack/dsh-workspace-digest
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install cui-stack/dsh-workspace-digest to install the plugin into your current profile; source code: https://github.com/cui-stack/dsh-workspace-digest
About this plugin
When a model needs to see what files exist in a workspace, it usually has no built-in way to do a simple directory listing. dsh-workspace-digest fills that gap: it registers a workspace_digest tool inside DeepSeek Harness so the model can request filenames by directory during inference, without any extra scripting on your side.
The plugin ships as an official bundle, installs into a profile with a single dsh command, and requires no --patch flag or native dependencies. The only tunable is maxEntries (default 20), which caps how many filenames are returned per call; you can override it per id in the profile cordis.patch.yml to fit workspaces of any size. Under the hood it is pure JavaScript, keeping the footprint minimal.
If your DeepSeek Harness workflows rely on directory awareness code repair, documentation Q&A, project summarization, or any pipeline where the model must locate files before acting this plugin gives you a ready-made, one-line tool that returns the current directory listing on demand.
Use Cases
- Listing workspace files so a model can locate targets during inference
- Surfacing directory structure in code repair or doc Q&A pipelines
- Letting a model see the file inventory before generating a project summary
Best For
- Developers building directory-aware inference flows in DeepSeek Harness
- Teams needing a zero-config file listing tool for model calls
- Users extending their toolset with model-inference category plugins
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.