dsh-plugin-canvas-pro
Run the following command in DeepSeek Harness:
dsh plugin install lokih1028/dsh-plugin-canvas-pro
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install lokih1028/dsh-plugin-canvas-pro in the DeepSeek Harness terminal to install the plugin; the full source is at https://github.com/lokih1028/dsh-plugin-canvas-pro .
About this plugin
When an agent in DeepSeek Harness produces an HTML page or a Mermaid diagram and you want to tweak one line, swap a color, or roll back a step, you are usually forced to regenerate the whole document. dsh-plugin-canvas-pro turns canvas artifacts into a first-class workspace for the agent: a Canvas button appears in the Web session header, a resizable drawer previews HTML, Markdown, Mermaid, SVG, and React source in a sandboxed iframe, and eight model tools let the agent create, patch by selection or line range, read the current content, inspect the latest text selection, list snapshots, diff versions, revert a past state, and export in multiple formats.
Three capabilities stand out. First, text selection is a first-class agent input: highlight a span in the preview to fire Explain, Refactor, Fix, or a custom prompt straight into the current chat, and the agent can also read that span via get_canvas_selection. Second, every patch becomes a snapshot; the agent can list, take a unified or split diff, and revert to a previous state without rewriting the file from scratch. Third, export is a model tool, not just a UI button: export_canvas_artifact returns Markdown, standalone HTML, SVG, PNG (with SVG fallback), or a JSON archive, all callable mid-conversation.
This plugin suits teams that build and iterate on rich-text artifacts inside DSH: sketching Mermaid architecture diagrams with targeted node edits, refining HTML prototypes through successive partial patches and version comparisons, or exporting the final result as a PNG for a deck. Artifacts persist locally under $DSH_HOME and survive Web restarts, with no external service dependency.
Use Cases
- After the agent generates a Mermaid diagram, the user selects a span and triggers Refactor for a targeted node edit
- Iterating on an HTML prototype by patching line ranges and taking a split diff between two versions
- Exporting the final artifact as a PNG for a deck or as a JSON archive into a downstream pipeline
Best For
- Developers who build and iterate on documents, diagrams, and prototypes inside DeepSeek Harness
- Technical writing or product teams that need the agent to patch spans instead of regenerating whole files
- Engineering users who require version history, diffs, and multi-format export of canvas artifacts
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.