DeepSeek Harness Plugins · Client
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
dsh-conversation-indicator is a client-side plugin for DeepSeek Harness Web GUI that provides a conversation indicator with a vertical capsule panel marking user messages for quick scrolling and navigation in long sessions.
A session-level Markdown writing pad for DeepSeek Harness Web, docked in the conversation sidebar for drafting, previewing, and AI local rewrites within a unified workflow.
dsh-launcher-lifetime plugin controls node service residency with three modes: always-on, tray resident, or follow window, to optimize memory usage and startup speed.
dsh-plugins-market is a creative marketplace for discovering, noise-filtering, trust-badging, and one-click installing DeepSeek Harness plugins.
This plugin controls iOS simulators via WebDriverAgent, enabling screenshots, taps, element lookups, and more without macOS accessibility permissions.
llm-adaptive plugin adds adaptive model routing to DeepSeek Harness, classifying each LLM request by complexity and routing to the matching backend provider.
DASH is a terminal TUI for the official DeepSeek dsh CLI, featuring native ANSI rendering, streaming sessions, and oh-my-pi-style keybindings and settings.
A kanban board plugin for DeepSeek Harness Web UI that organizes sessions into workflow columns for planning and triage with drag-and-drop functionality.
Adds a sidebar to the DeepSeek Harness Web GUI that indexes user conversation turns, with click-to-jump, scroll-synced highlighting, responsive design, and bilingual support.
An enhanced Agent Client Protocol server for DeepSeek Harness, built for Zed editors, offering streaming, telemetry, model control, and session management.
A matter-aware legal workspace plugin for DeepSeek Harness that provides a focused environment for litigation, arbitration, and other legal matters with document management, OCR, and task tracking.
dsh-tool-describe-image is a DSH plugin that enables DeepSeek to understand images via any OpenAI-compatible vision endpoint, converting them to text or HTML descriptions, with an interactive floating whale pet.