dsh-wsl-keepalive
Run the following command in DeepSeek Harness:
dsh plugin install TheColdWorld/dsh-wsl-keepalive
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install TheColdWorld/dsh-wsl-keepalive to install this plugin; the source repository is hosted at https://github.com/TheColdWorld/dsh-wsl-keepalive.
About this plugin
Running DSH WebUI inside WSL for local inference means Windows will quietly shut down your distribution after a few minutes of apparent idle time, killing your session mid-generation. dsh-wsl-keepalive exists to solve exactly that: a single toggle under Settings → Plugins that, when flipped, keeps the distro alive with zero scripting.
Under the hood it launches dbus-daemon inside the target distro via wsl.exe so Windows detects a live process and skips idle shutdown. Before starting it checks with pgrep to avoid duplicate launches; on stop it kills only the specific PID it recorded, leaving other dbus-daemon instances untouched. The settings panel shows real-time status, PID, and distro name at a glance. All three configurable fields (distro name, user, wsl.exe path) carry runtime validation—wrong values are reverted instantly with a red error message, and good values persist across restarts via a JSON config file. UI strings follow DSH global locale, switching between English and Chinese seamlessly.
If you rely on WSL for day-to-day local model inference and are tired of your distribution being silently reaped, this plugin is the lightest fix: one toggle, one flip, and your WSL session stays as long as you need it to.
Use Cases
- Keep DSH WebUI sessions alive during local model inference inside WSL
- Prevent Windows from auto-shutting down idle WSL distributions
- Toggle keepalive from the WebUI settings panel with a single click
Best For
- Developers running DSH WebUI for local inference inside WSL daily
- Users frustrated by Windows silently killing idle WSL distributions
- Anyone who wants script-free, GUI-driven WSL keepalive management
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.