DeepSeek Harness PluginManifest Valid

c-vision

Vision and computer-use plugin bundle for DeepSeek Harness providing window capture, OCR, and mouse/keyboard automation.

VisionTerminal & TUIDeveloper ToolsSkills & Workflows

0GitHub Stars0ForksUpdated2026-09-17

Source

Install

$ dsh plugin --profile tmp add github:cczzyy-cn/c-vision # 作为 bundle 自动挂载

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

c-vision (Vision) is a DeepSeek Harness plugin bundle that grants agents visual perception and desktop automation capabilities. By integrating a bundled Python cvision module with DSH attachment services and Cordis runtime tools, it enables agents to list open windows, capture screen or window images natively, run local OCR, and execute mouse/keyboard actions for closed-loop computer-use workflows.

Key Capabilities

Native window and full-screen image capture using Windows Graphics Capture (WGC) and macOS screencapture
Screen and window OCR text extraction with native system OCR or Tesseract fallbacks
Window enumeration via list_windows tool
Mouse automation including absolute coordinate single click, double click, move, and scroll
Keyboard automation including text typing and shortcut key press simulation
Window management and focusing via title substring matching
Automatic image fitting and resizing within DeepSeek Harness attachment size constraints

Useful For

Autonomous desktop task execution requiring vision-based feedback loops
Extracting text from running application windows and terminals without full image token overhead
Automating GUI workflows in browsers and desktop applications through mouse and keyboard actions
Inspecting background or occluded windows non-intrusively without stealing focus

Who It Fits

Developers building computer-use agents on DeepSeek Harness
Automation engineers running desktop GUI RPA tasks with DeepSeek models
Power users seeking multimodal visual assistance for local desktop applications

Documented Limitations

Requires local Python 3 environment and pip dependencies (requirements.txt) installed on the target machine
Linux backend implementation is currently placeholder/stubbed (Phase 2)
macOS requires explicit Screen Recording privacy permissions in System Settings
Direct mouse and keyboard automation modifies actual user hardware input and can cause misclicks if coordinates are incorrect

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.

No automated security signals have been published yet.

Source & Registry Notes

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

Source repository: cczzyy-cn/c-vision
Registry source: GitHub · dsh-plugin topic
Registry classification: Plugin Bundle
Source checked: c8a002f · 2026-09-20

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.