dsh-draft-polish

When drafting text, we often struggle with awkward phrasing or imprecise expressions, especially during chats or writing where quick optimization is needed. Traditional tools may be limited by system concurrency, preventing polishing while a session is active and disrupting the creative flow. The dsh-draft-polish plugin, designed for DeepSeek Harness, solves this pain point by instantly making your drafts clearer and more appropriate.

The core strength of this plugin lies in its one-click operation: click the ✨ button beside the input box to invoke LLM-powered automatic polishing of your draft. It works even during active sessions, allowing you to polish without interrupting your ongoing conversations. Additionally, it intelligently carries recent dialogue context to help the model better understand your intent, enhancing the quality of refinement. All operations are zero-contamination, creating no new sessions or histories—only modifying the input box draft—and results are isolated per session for accurate backfill.

dsh-draft-polish is suitable for a wide range of users, including content creators, writers, developers, and anyone needing to refine text daily. Whether you're quickly correcting expressions in a chat or enhancing fluency in document editing, it offers seamless support. By reusing existing model channels, it requires no extra configuration, providing a convenient polishing experience so you can focus more on the content itself.