DeepSeek Harness PluginManifest Valid

dsh-vision-opencode

Adds configurable vision-model image-to-text conversion, image analysis tools, and a vision-model selector to DeepSeek Harness.

VisionDeveloper ToolsSkills & Workflows

11GitHub Stars2ForksUpdated2026-08-20

Source

Install

$ dsh plugin --profile web add -w github:poiuyjie/dsh-vision-opencode

Plugin 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.

source-grounded

dsh-vision-opencode is a DeepSeek Harness web plugin for using a separate vision model with text-only main models. It can automatically convert chat images into text, provides the vision_read_image tool and vision-image-analysis skill, and adds Vision settings and a composer selector.

Key Capabilities

Automatically converts images sent in chat to text before forwarding them to a text-only main model.
Provides the vision_read_image tool for OCR and image understanding.
Provides the vision-image-analysis skill for OCR, chart, and screenshot analysis.
Adds a composer-bar selector listing image-capable models across configured providers.
Adds independent Vision settings for provider, model, automatic conversion, and reasoning policy.
Uses a 60-second timeout, one retry, and placeholder-text fallback for failed image conversion.
Leaves native multimodal main models on the native DSH image path.

Useful For

Use DeepSeek or another text-only main model while analyzing screenshots or uploaded images.
Extract text from images with OCR.
Analyze charts and screenshots through a separately configured vision model.
Choose a vision-capable model without changing the primary chat model.

Who It Fits

DeepSeek Harness web users with text-only primary models.
Users who want separate configuration for image analysis models.
Users needing OCR or screenshot and chart understanding in DSH chats.

Documented Limitations

Requires configuration of a vision provider and model before image conversion can be used.
Requires Node.js 20.3 or later.
Depends on compatible @deepseek-ai DSH peer packages in the >=0.1.0-rc.6 and <0.2.0 range.
Image conversion is limited to a 60-second timeout and one retry; exhausted failures become placeholder text.
Reasoning-disable behavior varies by provider and is not guaranteed, especially for forced disable mode.
The README warns that old image-containing conversations may no longer be sendable to text-only main models after uninstalling the plugin.

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

Dependency Counts

package.json declares 0 runtime, 0 development, 5 peer, and 0 optional dependencies.

info

Dsh Bundle Declared

package.json declares DSH bundle metadata.

info

License Declared

GitHub reports the repository license as MIT.

info

Package Manifest Available

A root package.json was captured and can be inspected by the registry.

info

Source Available

Public GitHub source metadata is available for this registry snapshot.

info

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository: poiuyjie/dsh-vision-opencode
Registry source: GitHub · dsh-plugin topic
Registry classification: Plugin
Source checked: a4ba066 · 2026-08-20

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.