Plugins

Learning mode that explains each agent step (WIP).

Encrypted local credentials vault for the Harness: a web settings page and vault_* tools to store, search and copy passwords, API keys, TOTP secrets and card data, with health audits, expiry rotation, imports/exports and read-only/ask access modes.

Per-call cache billing inside the context-meter popover: current step, round and session totals with hit/miss/output tokens and amounts, automatic Beijing working-hour peak/off-peak pricing, per-model rates, and third-party relays billed too.

Computer use through an external USB HID device: name a target in plain language, a vision model locates it in a cropped window screenshot, and a Raspberry Pi Pico 2 moves the real mouse and types on the real keyboard. 7 tools. Windows has window listing, per-window cropping and automatic window raising; macOS needs pyobjc and has neither; Linux is unsupported.

Session-bound persistent TypeScript REPL for DSH PTC mode: bindings and imports stay live across run_code calls, edit_run_code replaces one line without resending the block, module import/export syntax is adapted via AST rewriting, and durable bindings are restored from the session journal after a cold restart.

DSH Web GUI right sidebar: VS Code-style file explorer plus source control (git status, stage/unstage/discard, commit, diff, commit graph, fetch/pull/push sync, branch and remote management, git init, multi-repo switch, color-coded file status, file-type icons), an editor view that opens every file reached from the chat, the file tree or source control (Markdown/HTML/PDF/image previews, 15-language syntax highlighting, hunk-level diff editing, Ctrl+S save), file operations (new/move/copy/delete, reveal in file explorer), a collapsible icon rail, a multi-tab bottom terminal (ConPTY), and commit messages generated from the staged diff by a host-configured model.

AI-for-Science research workflow skill for DeepSeek Harness, aimed at researchers and grad students using agentic AI: multi-angle literature review with per-angle files, experiment design with a caveat list, figures and paper reading; each job returns files plus a cross-verification brief for a second model. Zero deps, MIT.

Zero-dependency, event-driven voice announcement plugin: no extra model, no token cost. Speaks with the system's built-in natural voice, supporting both Windows and macOS; final-reply announcements, approval & question alerts, optional event announcements (turn end, command done, goal change, tool errors, todo updates), replayable final replies, and a bilingual visual settings page.

Use wallpapers downloaded in Wallpaper Engine as the DeepSeek Harness web background: video wallpapers play in place, image wallpapers use the highest-resolution standalone artwork, with cover/contain fit modes and a 0-100 panel dimmer.

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.

Telegram mobile remote that attaches the phone to a live DeepSeek Harness Web session: workspace-then-session picker (/sessions), /last for the previous Q/A, /model switch, and native Rich Message replies.

Multi-engine web search for coding agents — one SearchBoost core (fused search, fetch, X) with MCP (Cursor, Codex, Claude Code, Grok Build, Antigravity), pi extension, and DeepSeek Harness bundle adapters

Role-based multi-agent team for DeepSeek Harness. The /team command assembles up to 12 specialist subagents (product, architect, researcher, UI, backend, frontend, database, security, reviewer, QA, DevOps, docs) and runs a nine-phase gated pipeline (clarify, research, design, spec review, plan approval, implement, review, test, deliver) over a shared workspace of artifacts (SPEC, PLAN, TASKS, STATE, REVIEW, TEST). Quality gates are enforced by plugin code rather than requested in prompts: an unfinished task cannot be marked completed, quality findings must be adjudicated by the reviewer or QA role, and coverage gaps or rework over budget appear live in the overlay and in /team status. Ships a live team overlay (phases, roster with per-role models, task DAG, artifact preview, decision buttons), a persistent-team mode that resumes across sessions, and per-run token and time accounting. Zero runtime dependencies, no build step.

Real mouse and keyboard for a DeepSeek Harness agent: describe what to click in words, a vision model finds it, a USB HID device clicks it.

Plugin governance: new/updated plugins stage in a candidate area first, then hot-mount after startup — success promotes, failure auto-disables, repeated failures freeze, one-click safe mode, sidebar diagnostics panel (npm: `dsh-my-guardian`).

Task reliability: auto-retry on model timeouts/failures, auto-continue unfinished tasks, completion verification, thinking-repeat interruption, resume after sleep or restart, autopilot mode and a remote trigger endpoint (npm: `dsh-task-reliability`).

Seven agent-preset modes for DeepSeek Harness, each with a fixed tool boundary, per-mode model routing and fast/deep omd_task delegation tiers, plus a plan-to-execute loop (/mode, /start-work, goal-backed /ulw) and a settings page that remaps models per mode and per tier.

"Single-Core Writing" plugin for DeepSeek Harness: Thoroughly tool-streamlined and output-hardened for quantized small models, suitable for running long-novel pipelines locally with on-device models.

Forces context compression when running low-context llama.cpp models such as Qwen3.8-27B, improving model efficiency and usage smoothness, together with well-refined compaction prompts.

UI harmonizer for DeepSeek Harness: normalizes unfinished or self-contradictory official UI, reconciles style conflicts between installed plugins, and unifies the visual language via official design tokens — with live-customizable chat width, markdown font size, workspace scale, UI font stack and a rounded center-column card from Settings → General; client-only dsh bundle, zero model cost, no host side, removable without residue.

A whip for DeepSeek V4 Pro: it watches the reasoning fingerprint of every thinking block and pulls the model back when it slips from focused "We will / I will" mode into scattered "let me" mode.

Per-plugin usage budgets and cost caps on the llm/stream waterfall: per-model usage accounting, session and monthly budgets with warn/block transitions, latency windows, and a carbon-footprint estimate.

GitHub Copilot LLM adapter: OAuth device-flow sign-in, live model discovery from the Copilot API, vision support for image-capable models (gpt-4.1, gpt-4o), and two wire protocols (Chat Completions and Responses API) with automatic endpoint routing per model.

OpenCode Go usage bar in the composer dock: rolling/weekly/monthly quota rings, DSH-session token & cost stats priced with live official rates, per-model and current-session scope breakdowns with live per-session updates, one-click API key copy.