AI Agent Hub
Back to plugins
🖥️

dsh-client-ui-code-diffs

Client Updated 2026.08.18

Run the following command in DeepSeek Harness:

dsh plugin install linhoulong123/dsh-client-ui-code-diffs

Paste the following prompt into your AI chat to install this plugin:

Install this plugin by running dsh plugin install linhoulong123/dsh-client-ui-code-diffs in DeepSeek Harness (source: https://github.com/linhoulong123/dsh-client-ui-code-diffs) and restart dsh web to activate it.

About this plugin

Every time a model runs an edit or write tool in a DeepSeek Harness conversation, the file change is tucked inside a tool card. The stock view lacks a clear red-green contrast, so after dozens of modifications in a single session it becomes painful to gauge the scope and correctness of each change at a glance. dsh-client-ui-code-diffs replaces the built-in card with a full-line inline diff: added lines carry a green background, removed lines a red one, and intra-line changes are highlighted at the character level.

The plugin ships syntax coloring for Kotlin, Java, TypeScript, JavaScript, Python, JSON, XML, and CSS, a header badge with line numbers and added/removed counts, and status labels that distinguish running, failed, and open-file states. Diffs exceeding 600 lines are automatically truncated in the middle to keep the message stream readable. It registers at a lower priority than the built-in card, so it silently takes over with no extra configuration. The entire implementation is client-side with zero server dependencies.

If you rely on DeepSeek Harness for code generation, refactoring, or batch edits and want to judge every modification right in the conversation flow without switching to a separate editor, this plugin is built for you.

Use Cases

  • Quickly reviewing every edit after a model performs multiple file changes
  • Distinguishing additions, deletions, and intra-line modifications in long conversations
  • Locating the scope of large file changes without overflowing the message stream

Best For

  • Developers using DeepSeek Harness for code generation and refactoring
  • Tech leads who frequently review model-made changes
  • Users who prefer inline diff views over switching to an external editor