Plugins

Manual context compaction for DeepSeek Harness: keep the latest N conversation items with stop or batch mode and a per-session compaction archive — embedded in the context meter popup when the meter panel slot exists, or a composer button otherwise.

Manual context compaction for DeepSeek Harness: keep the latest N conversation items with stop or batch mode and a per-session compaction archive — embedded in the context meter popup when the meter panel slot exists, or a composer button otherwise.

Official-grade DeepSeek Harness memory plugin backed by Hindsight (vectorize-io). Add /hindsight commands and model tools that recall, remember, inspect and forget through a Hindsight memory bank.

DSH plugin: per-model reasoning levels for custom API models, with family presets and an official-settings-page editor.

将 WorkBuddy 桌面 App 包含的模型自动接入 DeepSeek Harness — bring WorkBuddy desktop app models into DeepSeek Harness with zero configuration.

Model-facing MS Word (.docx) tools: read documents as Markdown or structured blocks, create documents from Markdown, and edit documents round-trip; ships binary fs providers (writeBytes, incl. a sandbox-preserving one) for hosts without it

Explicit, evidence-first multi-model consultation for DeepSeek Harness

Semantic model routing plugin for DeepSeek Harness

Bridges images into text for DeepSeek Harness: when the session's selected model cannot see images, a configured vision model describes them and the descriptions enter the durable session history as folded context rows — your message stays untouched.

Per-model reasoning effort (off/minimal/low/medium/high/xhigh) for GPT, Codex, Gemini and OpenAI-compatible models. Installs a helper, not a runtime adapter: run enable-capabilities.ps1 once per profile (Windows) and restart — it declares the levels in ~/.dsh/settings.yaml, with a disable script and a backup to revert.

Display companion for dsh-vision-router: keep Auto Vision routing, and put a square picture button on each original model row. Does nothing unless dsh-vision-router is installed.

DSH web plugin: a green/red status dot left of the composer model select showing whether DeepSeek API is currently in official peak time (red) or off-peak (green). Click for a live dashboard: countdown, 24h day track, stat tiles.

DeepSeek Harness plugin: a deployment deny list of filesystem roots the model may never touch - fences read, write, search, and shell tool calls at tools/pre-execute

Web UX kit for DeepSeek Harness: hover session actions and a large model picker, with more surface polish to come.

Live-editable Markdown notes in the DeepSeek Harness sidebar, with source mode, safe file CRUD, and autosave.

Specialist subagent delegation preset (orchestrator + 5 roles: oracle/designer/fixer/explorer/librarian) that seeds itself into your DSH home — background-first delegation, JSON-driven models/effort/tool permissions, scoped MCP for research.

A DeepSeek Harness plugin that records which files the agent edited, with which model, and in which session into git-ai

DeepSeek Harness model selector upgrade: one-level model & reasoning-effort picker with 5 Liang tiers, custom model aliases, and peak/off-peak price countdown.

Full-duplex voice mode for DeepSeek Harness: zipformer2 streaming ASR → editable draft, Edge TTS sentence-by-sentence read-aloud with live captions, true barge-in — on-device ASR, no API key. · DSH 语音双工对话：流式识别入草稿、按句朗读+实时字幕、开口即打断，识别本地推理、无需 API Key

将 DSH 聊天框模型选择改为先选择供应商、再选择模型。

A DeepSeek Harness plugin that records which files the agent edited, with which model, and in which session into git-ai

Route fresh DeepSeek Harness subagents to registered provider/model pairs declared in role Markdown files

Register AllRouter as a DSH provider: one base URL and one key for Kimi K3, Claude, GPT, Grok, GLM, DeepSeek, MiniMax-H3. Switch models without managing separate providers; every call reports real cost.

Periodic prompt injection for the DeepSeek Harness Web GUI: every N model invocations (conversation turns and tool-call steps both count) the plugin injects a user-configured prompt as a model-visible user message; the first injection happens at conversat