Plugins

Model-based permission approval: an approval-request answerer backed by a separate reviewer model.

Connect a local Obsidian vault to DeepSeek Harness: local Markdown indexing with keyword search by default and optional hybrid or semantic search through `obsidian_*` agent tools (hybrid and semantic need an OpenAI-compatible or Ollama embedding endpoint); agent access is off until you enable the conversation switch, and note writing is approval-gated, undoable and proposal-first; the index and vectors stay on your machine.

AIGC creation inside DeepSeek Harness: generate images, video, audio and 3D with open and closed-source models — parameter schemas, cost estimates, approval gates and a task panel.

A desktop dynamic island window for DSH (tested on Windows) that shows whether the agent is thinking, replying, running tools or waiting for approval; expanded, it shows plan progress, context and token usage, and lets you stop the task, approve permission requests and answer the agent's questions, with a separate usage window that estimates cost from DeepSeek's official prices. The island window needs an existing Electron runtime — reused from the dsh-orb plugin's downloaded runtime or pointed to by DSH_ISLAND_ELECTRON — and DSH Desktop 0.2.1-rc.5 on Windows; the package neither bundles nor downloads Electron.

Composer and sidebar helpers in one plugin, each switchable in Settings: a git Changes panel, Turn rollback, Resume after an AIProxy outage, Project memory (notes per project that later chats read, off by default), a Task board (kanban board of chats by their live state), Assistant (one lasting chat named Damon that sets reminders with dsh scheduled tasks and reads your other chats; it changes a chat only after an approval) with Assistant callouts, Assistant always free (long steps go to background subagents) and Assistant news from chats (off by default), Remote computers (another Mac's dsh over Tailscale, off by default), 11 extra UI languages and Automatic updates.

Connects DSH session status, approvals and questions to a native macOS notch app; requires the native app and Focus companion.

Apply configurable allow, deny, and approval rules to dsh tool calls, including shell command patterns.

Drive your local DSH from WeChat: a self-hosted iLink bridge over a persistent ACP session (context lives in DSH and is resumable), phone conversations grouped into the desktop workspace, with image recognition, file messages (Excel/Word/PDF, parsed by the agent itself) and voice transcripts. Tool approvals are answered inside WeChat itself (numbered card; ask/allow/reject policy; 120s timeout means reject), and any process can push a message to your phone through the outbox helper. Access defaults to strict (only registered devices are served); the phone channel's permission preset is configurable - see the README security section before sharing the bot.

Drive a local DeepSeek Harness agent from QQ and other IM channels through AstrBot: send tasks with /dsh, continue by quoting or @-mentioning the bot, answer approvals and ask_user_question prompts from chat (in parallel with the DSH web UI), and get results, inline images and file cards back. Ships session / workspace / model / permission commands, /compact and /last recovery, three progress-display modes, and group-shared or per-user session scoping.

Conversational Grafana operations: inspect live dashboard data across named sources, then make safe, reviewable dashboard changes with approval and version checks.

Local-first persistent memory for DeepSeek Harness: bounded pre-step hot-memory injection, exact and Chinese-bigram BM25 cold recall, six agent tools, approval-gated lease-lock writes with recovery receipts, and read-only governance and trajectory-review candidates.

Engineering workflow agent preset for dsh: five gated phases (requirements clarification, plan approval, TDD, parallel subagents, verified finishing) with six workflow skills adapted from obra/superpowers.

Biomimetic memory system for DeepSeek Harness: a single SQLite store (node:sqlite, WAL) whose Markdown tree is kept as a read-only mirror; memory tools, frozen-snapshot injection at session start, graded approval gating, on-demand extraction of the current session, near-duplicate merging, an archive and supersede lifecycle, memory metabolism (dream), memory pins, semantic recall over local embeddings, and the /memory command.

Chat with, monitor, and approve your DSH agents from WeChat via the iLink gateway: text both ways, session targeting, digest heartbeats, and numbered approval prompts.

Telegram sessions, topics, model controls and approvals for DeepSeek Harness.

Telegram bridge for DeepSeek Harness: forwards agent questions (with inline buttons), tool approval requests, turn summaries and subagent messages to a Telegram bot, and takes the answer from Telegram or the DSH Web UI, whichever comes first. Ships a loopback control panel; host-only, no client UI.

Keeps the PTC run_code deadline from firing while DSH waits for your answer, plan review or approval, so a question can stay open longer than the shipped 120 s budget.

Native workbench to select project sources, distill and test capability candidates, configure scoped Slack access, and review approvals and persistent tasks.

Side-panel Chrome and Firefox extension that lets DSH drive the tab you are on: pages reach the model as structured text with a numbered element list, and clicks, typing and navigation wait for your approval. Beyond upstream it adds viewport screenshots, opening a URL in a new window, scrolling inside nested scroll containers (chat transcripts, side panes, modal bodies), permission-mode, model and slash-command controls in the composer, and an opt-in developer-tools group (elements and CSS, page JavaScript, network requests) that is off by default and approved per call. The install is two halves: the bridge plugin comes from the prebuilt release tarball, and the companion extension ships as a tgz in the same release and is loaded unpacked, which is the only way Chrome accepts it. This project is a fork of omdsh-dev/dsh-browser: https://github.com/omdsh-dev/dsh-browser

Windows desktop control through UI Automation, window screenshots and mouse/keyboard input, with per-session app consent and action approvals.

Floating glass status bar for DSH on Windows 11 x64: an always-on-top capsule that mirrors session progress, the current step and pending approvals from a local DSH web profile, with peek and detail views; approval cards stay read-only until DSH returns its own decision receipt, and connecting to the official Desktop app is experimental and unverified.

Per-event sound alerts and system notifications for DeepSeek Harness. Four events — turn complete, question, approval and error — each get a switch, a sound picker with instant preview and a system-notification master switch. Twelve built-in tones plus a silent option are synthesized at runtime with the Web Audio API, so no audio files ship with the package; the four per-event defaults differ in register, timbre and rhythm (a high sine strike, a rising triangle pair, a repeated square pulse and a low sawtooth descent) so they are distinguishable without comparing pitches. A mode dropdown in General settings offers on, off and a smart mode that defers to the whale widget so the two never double up; the same row has a preview button and a gear opening a full settings dialog with volume, a skip-while-focused option and upload of your own audio up to 2 MB. A speaker button in the composer mutes sound without touching notifications. Works in the web GUI and the Electron desktop app.

Runs the agent's shell and file tools in disposable Podman containers, so its access is constrained to the workspace project by default, and adds container, image, volume and secret management with approval-gated control.

Windows Computer Use for DeepSeek Harness with 21 tools and verified targeting: controls are addressed through their UI Automation identity and patterns, and numbered marks are checked against the live window before input, so a stale target is refused instead of clicked. A known sequence runs in one computer_batch call, one model round trip; incremental observations, compact receipts (normally no more than about 6 KiB extra), bounded handoff waits (HANDOFF_WAIT_TIMEOUT/HANDOFF_HELD_TIMEOUT as continuations), the desktop-block cycle boundary and a deterministic 100 s owner-scoped leave backstop keep context small and prevent a stuck handoff from waiting out the tool deadline. Approval modes settle an unanswerable ask honestly: in a danger-full-access session with no approval card the action runs under that grant, every other sandbox — and every delegated subagent session — refuses and names the unavailable channel instead of reporting a user rejection, and gate diagnostics land in policy.jsonl. Human handoff, a model-visible stop brake and prebuilt native workers.