dsh-attachment-downscale
Run the following command in DeepSeek Harness:
dsh plugin install zoumutou/dsh-attachment-downscale
Paste the following prompt into your AI chat to install this plugin:
Install the plugin in DeepSeek Harness by running dsh plugin install zoumutou/dsh-attachment-downscale; the source repository is https://github.com/zoumutou/dsh-attachment-downscale
About this plugin
DSH image attachment admission rejects files that are over 2000px on one side, over 3.5MB, or over 40 million pixels, which makes uploading phone originals or high-resolution photos frustrating. dsh-attachment-downscale hooks into the attachments service entry point saveImages and uses sharp to automatically normalize oversized images before they are stored: proportional scaling keeps the long side at or below 2000px, and if the result is still too large, it steps down jpeg/webp quality or falls back to webp.
Animated multi-frame GIF and WebP files are skipped so their animation is not damaged, and any processing failure falls back to the original admission logic, preserving the existing rejection behavior without hiding errors. As a result, stored images always meet DSH limits, and later model requests, vision-bridge calls, and history exports receive compliant files.
It is useful for developers and agent-workflow users who regularly attach high-resolution phone photos, on-site images, or external visual assets to DSH, and who want fewer upload rejections, less manual compression, and safe handling of animated images.
Use Cases
- Auto-scaling phone photos or HD images before attachment admission
- Batch image processing in agent workflows to avoid rejections
- Skipping animated GIFs/WebPs safely without breaking frames
Best For
- DSH agent-workflow developers
- Users who frequently upload high-res photos to DSH
- Teams needing compliant images without manual compression
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.