AI Agent Hub
Back to plugins
🖥️

dsh-show-media

Client Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install NecromanAlbert/dsh-show-media

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

Clone the source repository at https://github.com/NecromanAlbert/dsh-show-media and run dsh plugin install NecromanAlbert/dsh-show-media, then restart the Desktop GUI to use the dsh_show_media tool in your conversation card.

About this plugin

Harness conversation cards cannot render local file:// or Windows absolute paths, so a quick glance at an image or short video in your workspace usually means routing through the model-vision channel-something designed for the model, not for the person actually staring at the screen. dsh-show-media fills that gap with a single lightweight tool: the agent calls dsh_show_media with a file_path, and the card renders the file or a video frame right in place. Clicking opens a preview showing the original file bytes, dismissed by Esc, a backdrop click, or the close button.

The plugin supports PNG, JPEG, WebP, and GIF images as well as MP4, WebM, MOV, M4V, and MKV videos (capped at 64 MiB), and always displays raw file bytes rather than the official attachment downsample. Relative paths resolve against the current session workspace. The preview is strictly scoped to the current conversation card-it does not claim shell.overlay and is not a second global gallery.

It is aimed at developers who frequently need a quick, human-oriented peek at local media files during a conversation. It is not a replacement for read_image, not a process-wide floating overlay, and not a multi-item gallery-just an extra for-humans preview slot inside the conversation card.

Use Cases

  • Quickly view a locally generated or referenced image in your workspace
  • Preview a short video clip without leaving the current conversation
  • Compare local media files side by side during a session

Best For

  • Developers who frequently work with local image or video files
  • Users who need a quick, human-oriented media preview during conversations
  • Developers who prefer a casual local-peek workflow over the model-vision path