Umans Vision Handoff
- Updated June 25, 2026
- karutoil/umans-claude-vision
umans-vision-handoff is a Claude Code plugin that adds a view_image tool so text-only Umans models can analyze images by delegating to native-vision Umans models.
About
umans-claude-vision is a Claude Code plugin that adds a view_image tool for Umans text-only models such as GLM 5.2. When the active model cannot see images, the plugin hands screenshots to a native-vision Umans model like Kimi and returns text analysis. Install it when you code with via-handoff Umans models but still need screenshot and diagram understanding.
- view_image tool for text-only Umans models marked via-handoff
- Routes images to native-vision Umans models with Kimi as first choice
- Supports GLM 5.2 and similar text-only providers on Umans gateway
- Lightweight Claude Code plugin from karutoil
Umans Vision Handoff by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add karutoil/umans-claude-visionAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| Last updated | June 25, 2026 |
|---|---|
| Repository | karutoil/umans-claude-vision ↗ |
How can I use screenshots and diagrams when my active Umans model is text-only and marked supports_vision via-handoff?
Route image reads to native-vision Umans models when the active text-only model cannot process screenshots or UI images.
Who is it for?
Claude Code users on Umans gateways running GLM 5.2 or other via-handoff text models who still need image understanding.
Skip if: Setups already using native-vision Umans models like Kimi K2.7 that accept images directly without handoff.
What you get
Images are analyzed by a vision-capable Umans model and returned as text the active model can use in the conversation.
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
FAQ
Which models need this plugin?
Text-only Umans models such as GLM 5.2 marked supports_vision via-handoff.
Which model analyzes the image?
A native-vision Umans model, preferring Kimi first with configured fallbacks.