dsh-cot-summerization
Run the following command in DeepSeek Harness:
dsh plugin install MeowLynxSea/dsh-cot-summerization
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, run the command: dsh plugin install https://github.com/MeowLynxSea/dsh-cot-summerization.
About this plugin
In the realm of open-source large language models, the full exposure of chain-of-thought (CoT) reasoning often leads to visual clutter and privacy concerns, whereas closed-source models typically hide these processes for a cleaner interaction. The dsh-cot-summerization plugin bridges this gap by intercepting the raw CoT output and using a lightweight summarizer model to rewrite it into a concise, natural-sounding version. This allows users to enjoy the transparency of open-source while gaining the "mystique" and tidy interface similar to closed-source offerings. At its core, the plugin features intelligent interception and rewriting capabilities. It performs real-time incremental summarization during streaming, supports various styles such as native, concise, or custom, and ensures coherence through advanced boundary protocols. All settings are open-source and configurable, from summary length to error handling, giving users full control to tailor the experience—even disabling the plugin entirely to revert to raw CoT. This not only enhances user experience but also maintains compatibility with complex workflows like Agent Loop, ensuring multi-turn reasoning remains unaffected. This plugin is ideal for developers and researchers using open-source models who seek a more polished interaction. If you want to reduce CoT distractions, protect privacy, or add closed-source-like elegance to open-source models, dsh-cot-summerization is a perfect choice. Whether for personal projects or enterprise environments, it offers a significant experience upgrade at minimal cost (only one additional small model call), effectively addressing the last gap in the open-source ecosystem.
Screenshots
Use Cases
- Hiding raw chain of thought in open-source model interactions for a cleaner UI.
- Providing compatible CoT summaries for Agent Loop to maintain multi-turn reasoning.
- Rewriting chain of thought via local models in privacy-sensitive scenarios to prevent data leaks.
Best For
- Developers using open-source LLMs who want to optimize CoT display.
- Researchers needing to maintain transparency while controlling information exposure.
- Enterprise users seeking auditable CoT hiding solutions.
Related Plugins
A service-oriented sidebar framework for DSH, offering a full workbench with file explorer, embedded browser, real terminal, Git panel, and extensible plugin services.
A beautiful, practical Claude Code-style TUI plugin with pixel whale top bar, flowing glow title, real-time status line, streaming thought expansion, time rewind, context progress bar and TPS gauge — zero core changes.
The plugin market for DeepSeek Harness: browse/search/one-click install community plugins and themes, with updates, backup/restore, hot disable, diagnostics, and AI fix.
A persistent whale widget on the DSH web UI that shows DeepSeek balance, today's usage, and per-turn cost, with drag-and-snap and sound effects.