DeepSeek Harness Plugins · Model Inference
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
A motor vehicle insurance claim decision tool that computes precise results from input parameters, with no network dependency.
A pure Node.js model inference plugin with no network dependencies, designed for lightweight offline operation.
A lighting power calculation tool that computes precise results from given parameters using industry experience estimates, built with pure Node.js and no network dependencies.
A model-inference plugin that lists gene editing future entries and retrieves their details by id or name, built in pure Node.js with no network dependencies.
A pure Node.js geometric mean calculator with zero network dependencies, built for model inference workflows.
A pure Node.js percentile calculation plugin with no network dependencies, designed for model inference workloads.
A pure Node.js model inference plugin with zero network dependencies, ready for offline use.
Returns heat-control cooking tips based on input parameters, a pure Node plugin with no network dependencies.
Let text-only models read images by intercepting the stream and substituting image blocks with a vision model's transcription, without switching models.
Manage LLM API request policies in DeepSeek Harness, including configurable request intervals, retry count overrides, and failure interception with retry or abort actions.
Displays the actual provider/model serving each subagent in real time within native tool-call rows, updating live on failover without modifying prompts or routing.
Injects a Solo pure-orchestration preset where the main agent only dispatches while explore, editor, verify, and general subagents form a closed-loop workflow, reducing context overhead.