dsh-prompt-forge
Run the following command in DeepSeek Harness:
dsh plugin install nina27486486/dsh-prompt-forge
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install nina27486486/dsh-prompt-forge in the DeepSeek Harness terminal to install the plugin; source code is available at https://github.com/nina27486486/dsh-prompt-forge .
About this plugin
Writing prompts often leaves you stuck with a draft that is technically correct but rough in structure and clarity. dsh-prompt-forge compresses the fix into a single action: finish your draft in the composer, tap the forge icon, and a streamed rewrite arrives with visible reasoning and far tighter instruction formatting.
The plugin is deliberately small (roughly 700 lines across host and client, zero runtime dependencies) and focuses on three things: a three-tier refinement mode (light polish, structured rewrite, expert-level task inference with variable placeholders), a plain-JSON vocabulary file (editable, git-friendly, tracks usage counts and tags), and a one-click export of any satisfying result into a dsh skill file for later slash-command reuse. Everything runs on the current session model, so no extra API key or endpoint configuration is needed.
Best suited for dsh users who write prompts frequently, want a zero-friction refinement loop, and prefer a tool with no database, no CDN, no telemetry, and a minimal supply-chain surface. All data lives inside the self-contained deployment folder.
Use Cases
- Refine a rough draft into a structured, instruction-clear prompt with a single click
- Store frequently used prompts in a plain-JSON vocabulary for instant insertion into the composer
- Export a polished result as a dsh skill file and trigger it later via a slash command
Best For
- Developers and product managers who write and iterate on prompts frequently in dsh
- Users who want zero extra setup and simply reuse the current session model
- Lightweight-tool users who prefer plain-file storage with no database or runtime dependencies
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.