Plugins

DSH plugin: auto-fill metadata (context length, output cap, display name, input modalities) for custom-provider models from the Model Metadata Registry (models.dev).

Terminal front door for DeepSeek Harness (dsh): an out-of-tree TUI mode bundle over dsh-base

Health checks for DSH plugins: manifest/patch/entry/build/pack/install verification, model-callable plugin_check, profile host-shadowing + BOM detection, environment diagnostics, and poison preflight.

Continual self-refinement loop: persistent memory, periodic review-and-refine, cross-session shared knowledge, and automatic rollback through a model-callable harness_refine tool.

Desktop notifications for run completion, model questions, and approval requests with per-platform system sounds, zero runtime dependencies.

Task board for DSH: create tasks with project and model assignment, run them manually or on cron schedules; new sessions in a project automatically pick up its todo tasks and move them to in-review when done.

MiniMax-powered multimodal plugin: real-time voice call mode (streaming conversation, floating dock UI), voice mode and mic voice input, plus image/video/music/speech generation and vision inspection tools.

Zero-dependency browser control: drive a real headless Edge/Chrome over CDP with 8 `pilot_*` tools (navigate/click/type/keys/eval/screenshot) plus a live draggable cockpit panel in the Web GUI - text-first snapshots for text-only models, no Playwright, no API key.

Optimize a raw instruction into a professional prompt through the harness llm service — four-section (Role / Task / Context / Format) or a headerless plain-text style. Context-aware by default: the recent conversation is injected as background and may enrich the ## Context section. Long outputs resume from the truncation point with auto-expanded budgets; validated results are cached (identical requests cost zero model calls); a unified call budget bounds worst-case cost. Composer ✨ one-click optimize, undo, cancel and a token-cost hint; prompt_optimize tool (optimize + iterate), auto-optimize hook, auto-detected zh/en role document.

Dedicated model routes, tools, and system guidance for vision, compaction, reviews, subagents, titles, and image generation.

Learning Mode agent preset: a coding agent that teaches while coding — concrete scenario-grounded explanations, Socratic guidance, and TODO(你) practice blanks, modeled on Claude Code's Learning output style; installable via dsh plugin add (dsh-learning-mode on npm).

Knowledge-base-driven agent preset: wiki long-term memory, a web KB UI, a management page and optional UI polish.

In-place conversation rewind in the same session window without forking (Claude Code /rewind semantics): a per-message ↶ button cuts the model context back to any user message, with an optional Claude-Code-style file restore from disk-persisted before-backups.

Codex App Server model provider using a local Codex login, with session reuse, Harness tool bridging, native action traces, and durable generated-image projection.

Simulates a Codex environment for selected model routes to improve GPT-family model tool-call success.

Use third-party subscription accounts in DeepSeek Harness: chat, image understanding, image generation, and image editing with the models your subscription covers, with available models and reasoning levels synced from your account; currently supports OpenAI/ChatGPT subscriptions, more providers planned.

Two-tier model routing: a strong tier plans, advises and reviews while a cheap tier implements, with plan-mode-aware auto routing, a high-impact escalation guard, failure auto-escalation, and subagent tiering.

Usage tiers and shareable stats cards: 🐟→🐳 ranks by monthly token burn with a locally-computed estimated percentile, precise pricing for 46 models across 6 vendors (including size-tiered Chinese pricing), backfill of pre-install sessions, and old-vs-new comparison across the 2026-08-17 rate change. Local-only, no telemetry.

Berührungslose Laufzeit-Telemetrie für DSH: jede Sitzung schreibt Aufgabe, Modellstufe, Tools, Fehler, Dauer und Status automatisch in ein menschenlesbares Journal mit Statistikbereich (reiner Beobachter — keine Tools, keine LLM-Aufrufe, keine Injektion).

Nutzt den lokalen Codex App Server als Modellanbieter in DeepSeek Harness.

Ein dsh-Plugin, mit dem die KI automatisch VLMs (multimodale Modelle) für die visuelle Analyse aufrufen kann.

Gehosteter, kostenloser Vision-Sidecar für DeepSeek Harness mit dauerhaften Sitzungsbelegen.

Ein HUD in der Ecke für DeepSeek Harness, das den Token-Druck der Sitzung, Kosten je Modell sowie Tages-/Monatsverbrauch auf einen Blick zeigt, mit editierbarem Budget und Guthaben, das deine Ausgaben mitverfolgt.

DeepSeek-Harness-LLM-Provider für Codex-Modelle über verwaltetes ChatGPT-OAuth.