プラグイン

Inspects ONNX models (inputs/outputs, shapes, opset, operators, size) and benchmarks them with ONNX Runtime on the local machine or on a remote target over SSH, reporting average/P50/P95/P99 latency, FPS and the execution provider actually used, with PASS/FAIL gates for FPS, P95 latency and model size. Requires Python 3.10+ with numpy, onnx, onnxruntime, and psutil installed (pip install numpy onnx onnxruntime psutil); remote targets need python3, numpy, and onnxruntime.

A small dock in the DSH Web UI: restart the host or reload the page in one click. Restart asks first (now / 30s / 60s / 2min / cancel); the dock drags to any edge only after a 320ms hold, and stays anchored when the window resizes. Refresh refuses to run unless the plugin bundle is verifiably loadable (timestamp baseline + real bundle fetch + syntax compile), so a reload can never brick the app. Restart is Windows-only (PowerShell 5.1); the UI and page refresh work on all platforms.

Command resolution forensics: which file a command actually runs, which PATH entry won, what it shadows, how many entries are duplicates or dead, and whether the copies behind one name report the same version.

Speaks a cue for DeepSeek Harness on Windows: a configurable phrase when an agent turn completes normally, a second phrase when the agent needs you to choose — a question, a plan approval, or a permission request — and a time-of-day greeting when a new conversation starts. A small bubble in the lower-left corner mutes and unmutes every cue with one click, with its own toggle chime. Phrases are synthesized with a Microsoft Edge neural voice (skipping aborted and errored turns, subagent turns, and rounds that a goal is still driving), cached under the DSH home directory and replayed locally, with the desktop SAPI voice as the offline fallback.

Windows desktop notifications with selectable sounds: five bundled sound packs (45 clips) plus imported audio, one sound per event, and clicking a notification opens the matching session.

Daily AI digest by email: fetches ten sources (OpenAI, Anthropic, Google DeepMind, Hugging Face, arXiv, Hacker News and others), clusters near-duplicate stories, scores them in code and with an LLM, then emails the top three on a daily schedule with a four-part reading (facts, why it matters, impact, open questions). Adds four conversation tools and a sidebar panel for previewing the digest, sending it now and changing the schedule.

Reports DeepSeek Harness session state (working, done, error) to the agent-pet macOS desktop overlay and lets it open sessions or start a chat.

Quiet break, water, long-session and late-night reminders delivered by three bundled chibi companions, with cloud bubbles and a light bounce. Works in-page without extra downloads; the optional macOS desktop window uses a separately downloaded Electron helper.

Toasts from DeepSeek Harness when a turn finishes, and when it stops for approval, a question or a plan exit. Native on Windows, text and timing configurable.

Turn-end reminders for DSH 0.2.x — the host folds each session's last completed turn into a projection and the desktop shell shows it as a real Windows toast, with per-outcome filters, background-only gating and a settings page. The Windows desktop shell must register its AUMID once before toasts appear.

Bidirectional GitHub repository sync for DSH: git sync with conflict rollback and pre-push checks, Actions/CI monitoring with per-step run detail, repository add/remove/rename, a web panel with local and remote file browsers, scheduled automation, and six agent tools.

Milestone progress soft-enforcement guard: injects persistent system-prompt rules and post-execute notices when tool calls omit visible progress text, with a Web settings form.

Registers native doubao_generate_image and doubao_image_status Agent tools to automate Doubao Web (doubao.com) via Tabbit Browser, automatically extracting watermark-free 2.7K wm0 original images into local files and multimodal chat attachments.

Extends the composer model selector menu with a third Context window row, preset capacities (64K–2M), custom input, provider-wide sync, and adaptive compaction guard.

Let external MCP agents delegate tasks to the existing DSH desktop, with visible sessions, user takeover, and optional approval handoff.

Generic monitor plugin for DSH Desktop: an HTTP polling rule engine. Poll any endpoint on schedule (multi-step, variable extraction, JS condition expressions); when conditions hit, wake a visible agent session to act — zero model calls otherwise. Configure everything in a sidebar GUI panel: create/edit/pause/test/fire-now/run history, same UX as the built-in scheduler plugin.

Dual-brain agent presets for DSH: one model plans the framework and delegates execution to a subagent pinned to a second model. Ships brain-ds (DeepSeek-V4.1-Flash plans, MiMo-V2.6-Flash executes) and brain-mimo (the reverse), switchable per session. Needs two providers configured (Xiaomi MiMo and a DeepSeek account), and the main brain must be selected manually.

Expose a running DeepSeek Harness instance over MCP, so an external agent can list, create and continue that instance's own sessions, delegate a turn and collect its result, answer the questions the DSH agent asks, and set how future approval requests resolve.

Codex-style hook management ported to dsh: hooks.json discovered from config, plugins and projects, each hook pinned to a sha256 fingerprint that must be approved before it runs; a review dialog in the composer, a `/hooks` browser, per-hook run stats, and trust / enable / dry run in Settings -> Hooks. The config file is dsh's own, with dsh event names.

Project Development Mode preset plus an in-chat development board for DeepSeek Harness: the lead conversation clarifies requirements and delegates to an Agent Team, and the floating board tracks each user change request through dispatch, execution, testing and acceptance with 7 board tools.

Self-maintained DSH web task board with a human acceptance gate, card kinds and editable notes on executed cards.

Scheduled tasks: run a prompt on a fixed interval, each run opening a fresh session, with a run history.

Provides a specification-driven engineering workbench with versioned requirements and documents, isolated Git workspaces, independent review, verification, and explicit human acceptance. Requires Node.js 24+ and Git.

Injects a very short conventions reminder (absolute INDEX path + version) once per session per version; when the conventions version changes, the next turn re-injects it (chat turns included). Optional hard guard: on operation-like turns, write/edit tools are rejected until the session has read the index (chat turns and shell are not blocked by default). Ships example checklists or points at your own directory; zero dependencies, configurable paths. Solves: conventions are written down but never pulled up when the work starts.