Plugins

A lightweight task-completion and approval notification plugin for DSH (dsh-notify-lite): when the agent finishes, it shows a Windows native Toast with the first paragraph of the latest reply (40 characters by default), plus taskbar unread badges, sounds and configurable notification options.

Let DSH speak: every reply is rewritten into natural spoken language before being read aloud, with code, tables, and Markdown handled gracefully. Works out of the box with system TTS; Doubao (third-party API) adds dialects and expressive voices; the local Kokoro engine runs offline to keep content on your device. Supports both DSH desktop and dsh web. Optional session sounds tell you when the agent is working, finished, or waiting for your answer — with swappable built-in tones and support for your own audio.

One MCP tool surface over two Windows desktop-automation backends (dsh-computer-use-win and windows-mcp). Both must be installed first: dsh-computer-use-win as a DSH plugin, windows-mcp via uv tool install windows-mcp --python 3.14, then run the bundled scripts/patch-backend.mjs once (it rewrites the local windows-mcp source; without it screenshot/read_webpage/launch_app fail) and disable the standalone wincu mount. It merges their 44 tools into 29 by dropping 10 duplicates, 4 dangerous ones (PowerShell, Registry, FileSystem, Process) and 1 tool-management entry, and blocks dangerous calls before forwarding through a reversibility-graded guard.

Long-conversation memory for DeepSeek Harness (DSH): keeps settled decisions, user corrections and exact numbers, with their conditions and sources, available across compaction and restarts in 100-1000+ turn sessions; one small budgeted evidence block per model step, with enabled skill rules re-applied in the same block.

Take over context compaction on both the manual and the automatic path, each runnable by a compaction model (falling back to a second model, or used directly) or by a mechanical fold that builds the summary card locally with no model call; the fold goes through the same official compaction transaction, so the replaced events stay in the append-only session log under their sequence numbers. It adapts code from dsh-compact-saviour and takes the mechanical fold from dsh-ledger-compact (both MIT) — install it instead of dsh-compact-saviour, not alongside it.

Connect Huawei CodeArts and OfficeAce models to DeepSeek Harness: it reuses existing logins (a logged-in CodeArts desktop client or dsh-codearts; the OfficeAce desktop app running and logged in on Windows), shows each account's remaining credits, toggles models per platform, and shapes requests like the official clients.

Runs the locally installed ZCode CLI as a `zcode` model route in DSH: all traffic is signed by the official binary, the plugin never sends its own request, forges client requests or sends credentials off the machine (the plan token is decrypted in memory only to hand back to the official CLI), no API key is needed, and both a paid one-shot and a free app-server channel are supported. Since v0.3.0 it ships an HTTP MCP bridge so tool calls run through the DSH approval surface. Backed by a contract document and a node --test gate.

Adds the models available in a signed-in WorkBuddy desktop app to DeepSeek Harness as a model provider, with the CN and international apps as separate groups. It is a rebuilt derivative of the npm package dsh-workbuddy-connect@0.7.9 (originally corrinehu/dsh-workbuddy-connect); it registers the same workbuddy / workbuddy-ai provider ids, so it cannot be installed alongside other WorkBuddy bridges.

Saves tokens in DeepSeek Harness sessions by reading any file by line range or symbol mention instead of whole, and by bounding oversized tool results while keeping a recovery path. Pruning happens at the source, not after content reaches the context. Measured 43.9% fewer prompt tokens in a modelled 30-step session.

VS Code style 44px icon rails for the side panels DSH collapses: the left rail keeps the sidebar navigation and the bottom action rows, the right rail lists the sidebarRightTabs ledger so third-party tabs appear on their own, and the right panel is drag-resizable between 300px and 60% of the window. Two switches under Settings → Activity Bar, saved in this browser.

Static-context audit for a DSH home: a six-layer baseline (rules, memory, skills, tool surface, sessions, KB), broken-path checks for a markdown capability map, memory duplicate/stale/over-limit reports, SKILL.md description compression, and tool-surface plus cold-start static-context estimates recovered from session logs. Four read-only DSH tools; writes require --apply and are backed up first. Requires Python 3.10+ on the host (stdlib only; all four tools shell out to the bundled CLI).

Read, message and live-follow local DeepSeek Harness sessions from inside an agent: four tools (session_list, session_read, session_send, session_chat) plus a matching Python CLI, over the desktop app local RPC; transcripts also read offline. Requires Python 3.10+; offline transcript reads also need the zstd CLI.

Native generation-time Mixture-of-Agents: every turn, reference models run the same session context in parallel as pre-run drafts and one aggregator executes with full tooling; config-as-model recipes, F mirror / R trimmed views, enforced subagent gating (dual-shape session indexing), derived per-recipe vision (vision refs relay image reads to text-only aggregators), tiered cadence with separate steering and patrol rosters plus frozen reuse between beats, degeneration guard dropping degenerate reference output, anonymized model face (failure notices carry no provider identity), and cache-friendly guidance placement (the rehearsal record is pinned where it was produced instead of appended to every request, so the cached prefix covers it and each step pays only for what was added since - on by default across every cadence).

Three things for DSH plugin development: reload the plugins whose files changed, restart the client when a reload cannot apply a change, and let the model trigger a reload itself. Reload uses DSH's own path, so untouched plugins keep running. Restart starts outside the client, so it completes even when the client is what gets replaced, and refuses while a session or job runs. Lightweight: no parameters, no added context, buttons off by default.

Drives the Claude desktop app from DSH over its claude:// deep links: open a Claude Code session with a prompt and folders, a Cowork task with files, or push a CLI session id back into the app. Also presses Ctrl+Enter for you (SendKeys) and reads the reply back from the local Claude Code transcripts under ~/.claude/projects/<cwd-slug>/<session>.jsonl — seven tools: desktop_send, desktop_continue, desktop_open, desktop_sessions, desktop_read, desktop_wait, desktop_probe.

Ships a minimal lesson ledger: brief.ps1 to look up past pitfalls, add-lesson.ps1 to record new ones, deployed into the skills directory and wired into AGENTS.md by the installer. The model checks it before working and turns what it hits into rules, so the same mistake is not made twice. This plugin watches whether that actually happens and shows the result in a tab and under the composer.

A memory panel for DeepSeek Harness with one store per workspace: persona traits, notes about you, dated events, and a first-person diary the model writes and cannot edit. Every entry keeps the quote it came from and can be retracted; workspaces can share one memory and be split apart again; optional per-person profiles for group chats; a capped event list with ★ marks. Local only, no network, zero dependencies.

A draggable task-tree panel inside the chat page: capture tasks by voice or text, drag to reorder, edge-snapping collapse, and a plain-text queue file you can edit or version yourself.

Composer infinity switch: while the highest reasoning effort is selected, one click appends a configurable notice to the system prompt.

Markdown display enhancements for the DSH Web chat, applied on top of the host renderer rather than replacing it: code blocks become cards with a language label and icon-only wrap and copy buttons, long code blocks, tables and Mermaid diagrams fold behind an expand control, Mermaid fences render as SVG with a diagram and code toggle, a fullscreen viewer with wheel zoom, drag and double-click to fit, and SVG or PNG export, tables get a sticky header and CSV or Markdown copy buttons, and a settings panel adjusts the diagram palette, code font size, the default diagram view and the three fold thresholds.

Action cards: the agent proposes a concrete action (create a file, edit a value, run a command), you edit it and apply it in one click, and the Host executes it and reports back.

Cursor system prompt reverse-engineered from Cursor.app, exposed as a selectable 'cursor' agent preset for DeepSeek Harness.

Work pet capability for DeepSeek Harness: a floating pet whose actions mirror what the current Session is doing, with swappable appearances contributed by other plugins.

A cockatiel appearance for the dsh-workpet capability in DeepSeek Harness.