DeepSeek Harness Plugins · Model Inference
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
Chat with a Harness Agent from Feishu or Lark via WebSocket long connection, reusing models, tools, and sessions.
Lets DeepSeek Harness agents drive ComfyUI to generate images and videos while managing workflows, assets, and job queues.
Use your ChatGPT/Codex subscription directly in DeepSeek Harness without an OpenAI API key, with models, quota, search, and image generation built in.
Register a native DSH LlmAdapter that streams free OpenCode Zen models with zero setup, no API key, no sidecar.
AI-native DSH quant plugin with 59 tools and an end-to-end research pipeline covering data, factors, ML, risk, and execution.
Brings Google Antigravity into DSH with multi-account pooling, dual quota monitoring, zero-paste OAuth login, and the official agy CLI.
Adds a CodeBuddy China provider to DeepSeek Harness, supporting WorkBuddy API keys and token login with model management in the WebUI.
AI simulates target users to walk through front-end interfaces, uncovering UX issues during development and providing reviewable optimization suggestions.
Adds a selectable smart routing mode to DeepSeek Harness that chooses automatic, inspect-first, or direct-execution guidance from the first user task.
Automatic provider/model fallback chains for dsh, switching routes on request failures with optional time-slot routing and role configuration.
Adds free image reading and image generation skills to DeepSeek Harness, with a vision model route and multi-engine failover.
Brings six OpenCode Zen free-tier models to DSH with zero config, no API key, quota-aware rotation and throttling.