DeepSeek Harness Plugins · Model Inference
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
Routes token-expensive requests to a local LM Studio model over its OpenAI-compatible endpoint, with auto-launch and context-window fitting.
Symbolic differential geometry via SymPy.diffgeom, supporting manifolds, tensor fields, exterior derivatives, wedge products, Lie derivatives, and curvature tensors.
Compress verbose voice-dictation text into token-efficient structured prompts locally, stripping fillers, repetitions, and politeness padding with zero LLM tokens.
A pure-Node color palette generator requiring no network or external services, invoked via its registered tools in a session.
A pure Node.js plugin for UUID generation and parsing that runs fully locally with no network or external service dependencies.
A pure Node.js text-reversal tool with no network calls or external services, executed entirely in your local session.
A pure Node placeholder text generator with no network dependencies; once installed, it registers a tool callable in sessions to produce lorem text.
Converts JSON data into serde Rust structs with a pure Node.js implementation via a single tool command.
A pure Node.js range generation tool for model inference with no network dependencies, ready to use after installation.
A pure Node.js prime number checker plugin with no network dependencies; invoke the registered tool in your session after installation.
A dependency-free array shuffle tool built with pure Node.js; invoke it in-session after installation to randomly reorder arrays.
A pure Node.js ellipsis and abbreviation plugin with no network dependencies, registered as a tool for in-session use.