Plugins

Notification outbox: agent proactively notifies via toast / Chinese TTS voice / sound effects (explosion, victory, alarm), 60s confirmation window voice-calls you back, volume boost, settings panel.

Open and edit Visio (.vsdx) drawings in a DSH panel: live SVG preview, shape and connector editing, layout checks, and a visio tool so the agent can read and modify the drawing. The core engine is Python standard library and needs no Visio or Office; requires Python 3.9+ on PATH, and Pillow for PNG export, snapshots and stencil previews.

HTML deliverable workbench in the file preview: viewport presets (desktop, laptop, tablet, mobile, A4), click-to-annotate elements with a notes/QA panel that docks to any edge or floats and sends the notes to the agent as one message, delivery QA grouped by category (figure/table numbering, orphans, clipping, overflow, page whitespace, WCAG contrast, images, dead anchors, banned words), PDF export through a local Chromium-based browser, and export of a self-contained annotated HTML that reviewers open in a browser to reply; also provides the `deliver_qa` and `deliver_export_pdf` tools and a `/deliver` command.

Adds an independent code font-size axis and per-plugin font-size sliders to the DSH font settings panel.

Project-level icon library: a right-sidebar Icons panel plus four agent tools (icon_list, icon_get, icon_save, icon_annotate) that turn SVG shapes into named, rule-described assets. Artwork can be pasted, imported from Iconify or IconMCP, adopted from a workspace scan, or generated; editing one drawing rewrites every code reference to it.

Multi-agent software engineering and security audit workflow with 8 role presets (Captain, CEO, Designer, Eng Manager, Reviewer/CSO, Builder, QA Lead, Release), 4-Wave DAG scheduling across 6 default concurrency slots (1-12), PostgreSQL Wire Protocol v3 + SQLite WAL fact store, 29 skills, and a docked right-panel web console. A fork of Jueze-2019/dsh-redteam-mode that keeps its package name, mount ids and redteam_* tool names, so the two cannot be installed together. The fact store uses PostgreSQL when REDTEAM_PG_URL, DSH_PG_URL, DATABASE_URL or PGHOST is set in the environment and a local SQLite file under $DSH_HOME/redteam otherwise; role presets are written to $DSH_HOME/.agent-presets/redteam.

Git panel for the DSH right sidebar: staging, commits, branch switching, fetch/pull/push, worktrees, a Cursor-style commit graph and GitHub pull requests (the GitHub card needs a GitHub token). It drives the system git binary, so the panel and the terminal always agree.

Suiji Word Book: a word book panel that shows one word per page with picture, phonetics, pronunciation, five phrases and an example sentence, plus fill-in-the-letters spelling practice and a 12-type word-recognition quiz. Word data is kept as plain files in a local directory, and an included skill lets an AI add cards from a word list. The “+Word” button (hidden until enabled in Settings) starts an AI session and by default switches it to danger-full-access; turn that off in Settings.

Picks each subagent model from the local catalog by task category, unless the caller already named one. Free and quota tiers are ranked together by per-model capability tags; a pricier model is used only when no cheap model is tagged as good at that kind of work. Only providers whose remaining balance is readable, or that are account-backed, enter the candidate pool; the rest stay listed but are not routed. A panel adds per-model manual priority above the automatic order plus allow/revoke controls.

Mistral subscription for DeepSeek Harness: model route, multi-account failover, and a composer usage panel.

Registers FreeLLMAPI as a dsh model provider, with the router itself bundled in the plugin. It launches the vendored server with the host binary and exposes the freellmapi route; the sidebar panel starts and stops that process, lists the router catalogue, probes which claimed models actually answer, and ranks them by the stability, speed and intelligence scores the router computes for itself. Provider keys, routing strategy and usage remain in the router console, which the panel links to.

Estimates what each conversation costs, with built-in official prices from 16 vendors (DeepSeek peak/off-peak hours and long-context tiers included), a breakdown panel with a currency picker under the composer, and each conversation's cost in the sidebar.

The mobile client for DeepSeek Harness: use the DSH running on your computer from your phone. The Android app signs in with the same account, picks this computer from the device list and opens its DeepSeek Harness UI on mobile. The desktop plugin handles registration and the connection, and nothing on the computer needs an inbound port. The UI adapts on its own: on a phone it becomes a touch-friendly layout, and returns to the desktop layout when a foldable opens or the window gets bigger. A PC-side panel shows login state, device quota, device list and the app download QR code.

On-demand proxy for DeepSeek Harness that leaves the system proxy, WinHTTP and TUN alone: its own Clash-style engine parses vmess, vless, ss, trojan, socks5 and http subscriptions and runs a rule engine (DOMAIN, DOMAIN-SUFFIX, DOMAIN-KEYWORD, IP-CIDR, IP-CIDR6, GEOIP with DIRECT/PROXY/REJECT) so only the requests that are slow when dialled directly take the proxy on the web, search, scripts and browser channels, with per-node latency and bandwidth tests, spoofing options and a loopback panel behind a per-instance token.

Breaks a STEM textbook into a four-level card map (book, chapter, section, knowledge point) in a sidebar panel, where each point names the defect of the previous approach it fixes and each chapter keeps its remaining gap as the motive for the next; marks key points with a star, shows theorems as cards with statement, conditions and symbol table, folds step-by-step derivations, and exports the same map as a self-contained offline HTML file, Word or PDF. Ships a textbook-parser skill and a textbook_archive tool (save_book / save_chapter / list_books / get_book / delete_book) plus a bundled 12-chapter Signals and Systems sample with 116 sections, 263 points and 601 worked examples. No runtime dependencies and no network access.

Adds the Miaoshou (Meituan CatPaw) desktop client as a DeepSeek Harness model provider. The 12 models mirror the ones the Miaoshou app picker shows, each carrying its credit-cost multiplier; thinking levels are translated into the upstream modelParams; a read-only panel reports available credits and the current plan. Sign-in is a QR link, so the desktop client does not have to be running.

Immersive weather theme for DeepSeek Harness: physical rain with splashes and ripples, snow in two crystal habits, a hearth, lightning with distance-coded thunder, a night mode, and a compact control panel.

Adds a strip of saved prompts above the composer: click a chip to insert its text at the draft caret, and manage the list — add, edit, reorder, delete — in an inline panel that opens in place. Prompts persist per profile and are also exposed to the agent as a quick_prompts tool; a fresh install starts with no entries.

Month calendar panel in the DSH main column: add and edit a day's entries inline, overlay the Host's scheduled reminders for today and tomorrow coloured by run state, and merge iCalendar subscriptions with the calendars already synced into this Mac's Calendar.app; a calendar_board tool exposes the same entries to any agent session.

UI plugin for DeepSeek Harness with a persistent customizable left Dock, cross-workspace pinned chats, a draft-first home, global search, a completion inbox and right-panel video previews.

Guarded decision reuse for the DeepSeek Harness: the model authors already-decided tool actions once, and receipt-supported ones execute through the native tools without another reasoning request. Ships a read-only decision-reuse panel with per-number dat

A second brain for DeepSeek Harness where Markdown is the database and git is the sync, with zero runtime dependencies. Core memory is injected into every system prompt, one file per note with weighted search; cards produced in the sidebar web chat land in an inbox first and are approved one by one as cloud notes or local-only notes (private/ is permanently gitignored and never pushed); web-chat progress is saved to progress.md with the last 20 drafts archived. A four-step flow lives behind one header icon: open web chat, extract cards, review them, back up to the cloud. The extraction prompt is editable in the panel and applies immediately.

Right-sidebar GitHub panel with four rising-repository lists — created within 2 days, 7 days, 30 days, and 20k+ stars still pushed — where the first three are ranked by star velocity (stars per day) rather than total stars, so repositories that gained traction fast surface above older ones with more stars. Each row shows the repository name, its own description, language, the stars-per-day rate and up to two topics, and links to the repository. A keyword search box queries the GitHub Search API directly, debounced and IME-aware so a Chinese or Japanese composition costs one request instead of one per keystroke. Reads the public GitHub Search API anonymously: no token, no login, no account, nothing else contacted. Lists cache for 30 minutes and search results for 5; outbound requests go direct first and fall back to a listening local proxy port only if direct fails.

An IDE inside DeepSeek Harness with a file tree, multi-tab editor, AI assistant with model picker and API-key setup, Git panel, PTY terminal, and Python/TypeScript LSP.