dsh-subagent-vision
Enables text-only DeepSeek sessions to process images by delegating vision tasks to a configured subagent with automatic paste-to-path conversion.
Install
$ dsh plugin --profile web add dsh-subagent-visionPlugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Registers the subagent_vision tool to delegate image reading tasks to a designated vision model
- Provides send-time image-to-path conversion for text-only chat sessions to prevent image admission errors
- Supports native drag-and-drop and paste intake with composer thumbnail previews
- Offers dynamic vision route configuration and one-click input modality declarations via harness settings
- Enforces magic-byte verification and a 25 MB upload cap for uploaded temporary image files
Useful For
- Reading screenshots or diagrams inside a text-only DeepSeek conversation without manual copy-pasting
- Extracting text, code, or structured summaries from pasted images using specialized vision models like Qwen
- Seamless multimodal query handling in DSH Web UI without session resets or manual model switching
Who It Fits
- DeepSeek Harness web profile users operating text-only primary LLMs
- Developers needing programmatic and interactive image analysis delegation within DSH
Documented Limitations
- Parent model only receives textual interpretations from the subagent and never sees the raw image data directly
- Vision subagents are one-shot and settled; continuous multi-turn dialogue with the subagent is not supported directly through this tool
- Requires a vision-capable secondary model configured in DeepSeek Harness settings
- Pasted images create temporary files on the host disk that require OS-level cleanup
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
package.json declares 1 runtime, 1 development, 7 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
GitHub reports the repository license as MIT.
A root package.json was captured and can be inspected by the registry.
Public GitHub source metadata is available for this registry snapshot.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- niuniuaba/dsh-subagent-vision
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin Bundle
- Source checked
- 9c02042 · 2026-09-19
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 0GitHub stars
- Forks
- 0
- Open issues
- 0
- Last commit
- 2026-09-16
- Last release
- No release detected
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Vision routing and pixel-level tools for text-only DeepSeek Harness agents with built-in free fallback.
View plugin →Persistent multi-model agent teams with dynamic DAG orchestration, run monitoring, and token insights for DeepSeek Harness.
View plugin →Session lifecycle manager for DeepSeek Harness with search, restore, idle cleanup, and diagnosis.
View plugin →Full office suite integration for DeepSeek Harness supporting spreadsheets, docs, slides, bases, and canvases.
View plugin →