Plugins
Explora, filtra e instala plugins de DeepSeek-Harness.
1219 plugins encontrados
API-MRI
lee-eee-eee/api-mri
Inspect API latency, streaming output and model-response fingerprints from a native DSH panel, with classic-task previews and public benchmark rankings.
dsh-memory
rosesssss66/dsh-memory
Local long-term memory for DSH: extracts preferences and project facts from sessions, injects the most relevant ones before each turn using local SQLite FTS5 search, and supersedes outdated facts. Requires Node 24+; the offline self-evolution loop needs an OpenAI-compatible LLM endpoint and API key.
dsh-card-tree
2031814001yuyue-tech/dsh-card-tree
Grows a conversation into an editable card tree where each card is one self-contained question and answer, with child cards, same-level branches and in-card follow-ups, and an overview for dragging, re-hanging, cross-referencing or pruning the graph.
dsh-llm-vertex
maryhodges5070-art/dsh-llm-vertex
Google Cloud Vertex AI (Gemini) provider: ADC or service-account credentials, project and region settings, thought-signature replay for tool calls, rate limiting, and a settings page with a live connectivity test.
dsh-tailtransit
jetruan/dsh-tailtransit
Runs a userspace tailscaled for Harness so web fetch, shell, plugin and LLM traffic can each use a Tailscale exit node, leaving routes and DNS untouched. Needs Tailscale's tailscaled installed separately; it starts a second per-user daemon, routes the selected traffic through a loopback proxy and looks up the exit IP through ifconfig.me. An auth key, if you set one, is kept in the plugin config.
alice-memory-llm
cycocyco/alice-memory-llm
Long-term memory for DSH where an LLM decides what to keep: turns pool while the session is idle, then the main chat model distills them into one-line entries in a single batched call; vector dedup runs before insert, so near-duplicates and saturated topics are dropped; recall uses hybrid vector plus FTS5 search over tiered hot, warm and cold storage; ships remember, recall, forget and prune tools. Zero dependencies, Node only.
dsh-galatea (message-channels)
loveandbeepbrother/dsh-galatea
Chat with your Galatea assistant from QQ (official bot) and WeChat: the same conversation, persona and memory as on the desktop, only the paired owner's private chat is accepted, replies go back where you wrote, and more chat apps can be added. Requires the Galatea framework.
dsh-plugin-cas-kb
niuniu-869/dsh-plugin-cas-kb
Chinese accounting and tax knowledge base for DSH agents, served by the hosted Quaesto service at api.accountingllm.site (needs network access; no account or API key, but your queries leave the machine): answers questions on CAS enterprise accounting standards, the small-enterprise accounting standards, CSA auditing standards and tax law with original articles, decision procedures, applicability checks and tax-accounting differences, exposed through 10 MCP tools and a china-accounting-standards skill.
dsh-galatea (framework)
loveandbeepbrother/dsh-galatea
Galatea: shape an AI assistant of your own. One assistant with its own persona and long-term memory (git-versioned plain files); long chats hand off to a fresh conversation without forgetting, and it tidies its memory in the background while you are away. It runs on whatever model DSH is set to and keeps its memory when you switch; every module can be turned off, and other plugins can extend it. Phone chat (QQ, WeChat) and Claude Code sessions come from the separate Galatea channels and agent-sdk plugins. Preview; needs DSH 0.2.2-rc.1 or a later 0.2.x.
dsh-one-gateway
qinyongliang/dsh-one-gateway
OpenAI-compatible gateway for every connected model: groups with failover, aliases, session stickiness, usage charts, API keys and a LAN listener (open to the whole network until you create an API key).
dsh-galatea (agent-sdk)
loveandbeepbrother/dsh-galatea
Run DSH sessions on the local Claude Code (Opus, Sonnet), with Galatea's memory and tools. Pick an auth mode in the profile (Anthropic API key recommended, or your existing Claude Code login). Requires the Galatea framework.
dsh-nextsteps
klaus-2/dsh-nextsteps
Suggested next-step prompts under every finished agent turn in the DeepSeek Harness chat: the model proposes a few followup prompts at the end of each turn, they render as clickable cards under that turn, and clicking one sends it verbatim as your next message. Each card stays attached to the turn that produced it across reloads and restarts, the generation language can be chosen from 15 languages or follow the conversation, and a settings page in the dsh settings covers the language, how many suggestions per turn, the header and the in-app update check. Requires DeepSeek Harness 0.2.x (web profile) and Node.js 20+.
dsh-kb-sync
xinming-chenniang/dsh-kb-sync
Pulls your chat.deepseek.com conversations into a local knowledge base — one Markdown file per conversation, a JSONL transcript and a catalog — and searches them from a control page inside DSH. Daily increments plus a full-history backfill from the official export, an integrity check, and an optional model step that runs through the host llm service and writes only after its JSON passes validation. Reads the web userToken from the DSH credential store and calls chat.deepseek.com only; no install scripts, no runtime dependencies.
dsh-laci-engine
atlantis-la/dsh-laci-engine
Local model scheduler for DSH on Windows 10/11 with DSH Desktop 0.2.x: a bundled console for engine/model import, VRAM estimation, load/close, and an OpenAI-compatible gateway.
dsh-rag (plugin)
fqh22/dsh-rag
Memoria duradera por proyecto para DeepSeek Harness: panel para explorar, buscar y editar conocimientos e indicación tras el primer mensaje para leer el paquete de relevo. El almacén es un servidor MCP local (Python + SQLite) del mismo repositorio. Instala el plugin por npm, clona el repo y configura MCP según README; localiza la copia mediante DSH_RAG_ROOT o rutas comunes. Los paquetes ponen la base estable primero y añaden hechos al final para reutilizar la caché de prompts.
dsh-ocg-session-head
verneuil/dsh-ocg-session-head
Inyecta x-opencode-session por conversación en rutas OpenCode Go registradas con un proveedor personalizado, mediante el waterfall oficial llm-pi-ai/request-headers. Sin dependencias ni parches de fetch; nunca sobrescribe una cabecera ya definida.
dsh-wechat-queue
steelunion/dsh-wechat-queue
Canal WeChat con cola secuencial persistente: guarda primero los mensajes en disco, responde de uno en uno y conserva el trabajo pendiente tras reiniciar.
dsh-system-one-plugin
xunzhuo/dsh-system-one-plugin
System Duo para DeepSeek Harness: combina pensamiento rápido/lento. Modelos de decisión System One clasifican, vigilan, protegen, enmascaran y revisan el bucle junto a System Two (vllm-sr/auto), con clave gratuita app.vllm-sr.ai.
dsh-budget-guard-plugin
existence001100/dsh-budget-guard-plugin
Plugin DSH que impone topes por tarea en CNY o tokens, con bloqueo automático para evitar costes descontrolados.
stockproof (dsh-plugin)
taoqizhang55-code/stockproof
Herramientas de análisis bursátil para EE. UU., acciones A de China y Hong Kong mediante el servidor MCP stockproof (Python 3.10+; instala con pip install "stockproof\[mcp,cn]"; los datos A/HK requieren Python 3.11+). Incluye noticias, resultados y anuncios con fuentes, explicación de movimientos de precios, filtro de todo el mercado A, laboratorio cuantitativo para probar estrategias con datos históricos puntuales y conjunto de prueba bloqueado, revisión de cartera y panel web local; 23 herramientas.
dsh-local-delegate
mythgreenland/dsh-local-delegate
Delega tareas mecánicas de texto (traducir, resumir, extraer, clasificar y reescribir) a un endpoint nativo de Ollama o compatible con OpenAI mediante rutas de archivo, para que el texto fuente no entre en el contexto del modelo principal. Registra local_transform, local_batch y local_health. Los niveles de IA local por Workspace (none/light/default/high/max) añaden el subagente local_worker y un modo solo de planificación. La opción SlaveAgentTeams detecta rutas de modelos montadas, ejecuta agentes hijos en paralelo con conmutación por fallo por nodo y registra la cadena de endpoints como rutas llm-pi-ai. Se configura desde la Web GUI y requiere un endpoint accesible.
dsh-plugin-sessiongraph
cc19990113/dsh-plugin-sessiongraph
Paquete para la capacidad sessiongraph: instálalo para añadir la búsqueda local del historial de conversaciones a un perfil de DeepSeek Harness.
dsh-fine-tune
songrui111/dsh-fine-tune
Ajuste fino local LoRA y QLoRA en un panel: importa JSONL, muestra una estimación de VRAM antes de iniciar el entrenamiento, evalúa y exporta un adaptador. Configura un entorno virtual de Python con un clic, pero debes proporcionar Python 3.10+ y una GPU NVIDIA de 6 GB o más. Verificado de extremo a extremo en Windows 11 con RTX 3060 de 6 GB; el instalador rechaza equipos que no sean Windows, por lo que Linux, macOS y otras GPU no están admitidos por ahora.
dsh-llm-minimax
elfgzp/dsh-llm-minimax
Añade modelos de la plataforma MiniMax a la lista de modelos de DSH: pega la clave API en Settings para poder seleccionarlos; las ventanas de cuota del M Plan aparecen junto al compositor.