DeepSeek Harness Plugins
Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.
A pixel-art and sprite-animation editor plugin for DeepSeek Harness, compatible with Aseprite project files, featuring drawing tools, AI integration, blueprint workflows, layers, and animation management.
A reliable todo panel plugin for DeepSeek Harness web that survives restarts and verifies completion based on evidence paths.
Browser MCP is a local MCP server that enables any MCP-supporting AI assistant to crawl and interact with any website using your real Chrome login state, including automated operations in the background.
An independently publishable DSH bundle that adds a categorized plugin manager to Settings → Plugins, supporting market search, installation, updates, and restart prompts.
dsh-grayprint is a dsh web plugin that analyzes paragraph openings in reasoning blocks to evaluate in real-time whether the thinking text resembles grayprint or current version samples.
An external vision proxy plugin for DeepSeek Harness that enables text-only LLMs to inspect, analyze, and understand images via OpenAI-compatible vision model APIs.
Adds Feishu OAuth login to DeepSeek Harness Web GUI for multi-user session isolation, allowing users to access only their own conversations.
A DSH configuration migration tool for exporting, packaging, and restoring configs across machines with verification and rollback.
The dsh-codex-workflow plugin enables Codex to perform read-only planning and review roles while DSH remains the sole executor, offering two workflow modes with human-readable audits and safe conflict resolution.
Enable using ChatGPT Plus/Pro subscriptions to access OpenAI Codex models in DeepSeek Harness via OAuth.
A DSH web plugin for usage statistics, featuring token trends, activity heatmaps, cache hit rate curves, and model usage breakdowns, all with hand-drawn SVG charts and theme adaptation.
dsh-route-resilience provides multi-route high availability, fault isolation, and observability for DeepSeek Harness, grouping provider routes for load balancing and automatic failover to ensure continuous model inference.