dsh-tool-vision-read
A DeepSeek Harness tool plugin that lets text-only agents describe local images through a separately configured vision-model route.
Install
$ dsh plugin --profile web add github:Mappedinfo/dsh-tool-vision-readPlugin 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 `vision_read(file_path, focus?)` model-facing tool.
- Routes image-description requests to a dedicated configured vision provider and model.
- Supports direct single-call execution and an in-process subagent mode.
- Commits image files through the DSH attachment service before vision-model processing.
- Supports configurable tool name, prompt, maximum image bytes, and maximum output tokens.
- Installs through the DSH plugin manager as a Profile Bundle and mounts via `cordis.patch.yml`.
Useful For
- Give a text-only agent a way to inspect and describe images.
- Route image understanding to a specialized vision model while retaining a separate chat model.
- Request focused descriptions of screenshots, photos, or supported image assets from a workspace.
Who It Fits
- DeepSeek Harness users with a configured vision-capable model route.
- Users running text-only agents that need image-description capability.
- Developers configuring DSH Profile Bundles and Cordis patches.
Documented Limitations
- Requires a DeepSeek Harness deployment and a configured provider route that declares image input.
- Only accepts PNG, JPEG, WebP, and GIF paths resolved relative to the calling session workspace.
- The default direct mode makes one vision-model `llm.stream` call; subagent mode uses a fuller agent loop.
- Provider and model configuration are required; configuration parsing fails if either route identifier is omitted.
- Image size is constrained by attachment limits unless `maxImageBytes` is configured.
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, 0 development, 8 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
- Mappedinfo/dsh-tool-vision-read
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 921d4af · 2026-09-14
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
- 1GitHub stars
- Forks
- 0
- Open issues
- 0
- Last commit
- 2026-08-21
- 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 visual tool suite for DeepSeek Harness agents with built-in free fallbacks.
View plugin →DeepSeek Harness plugin integrating ego-lite Chromium browser with 32 structured automation tools and live interactive screencast panel.
View plugin →Persistent multi-model Agent teams and bounded DAG workflows with GUI controls for DeepSeek Harness.
View plugin →Local-first AI agent runtime and DSH plugin bundle providing sandboxed sessions, MCP tools, audit logs, and session replay.
View plugin →