dsh-reasoning-effort
Run the following command in DeepSeek Harness:
dsh plugin install statem-li/dsh-reasoning-effort
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install statem-li/dsh-reasoning-effort in DeepSeek Harness to install this plugin; the source code is available at https://github.com/statem-li/dsh-reasoning-effort.
About this plugin
When running inference through DeepSeek Harness, reasoning effort is often the key knob that trades output quality against latency. Different tasks call for very different depths of reasoning, and the absence of a single configuration point can make the tuning process fragmentary.
dsh-reasoning-effort plugs into the Harness pipeline and brings reasoning-effort configuration into a manageable, plugin-level setting, so developers can control this parameter in one place rather than scattering hardcoded values across their codebase.
It is aimed at developers and teams who already build inference applications on DeepSeek Harness and want finer-grained control over reasoning intensity, particularly in services that need to handle requests of varying complexity.
Use Cases
- Configure reasoning effort centrally in the DeepSeek Harness inference pipeline
- Dynamically adjust reasoning depth for requests of varying complexity in a single service
- Compare output quality across different reasoning-effort levels during A/B testing
Best For
- Developers building inference applications on DeepSeek Harness
- Technical teams needing fine-grained reasoning-intensity control to balance quality and latency
- Engineers extending inference configuration within the Harness plugin ecosystem
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.