Plugins

Durable task-decomposition plans for DeepSeek Harness with a self-learning loop: every department subtask runs in three ordered phases (plan, execute, verify), the verify phase must submit a retrospective of problems, detours and resolutions, and that retrospective is written to a permanent lessons file, ranked against the new subtask and injected into later dispatches so the same mistake is not repeated. Adds a read-only dsh_workplan_lessons tool (11 model tools in total), a progress panel drawn as a ring, a stacked bar and per-stage steps, and a public ISO/IEC 15408:2026 security self-assessment with zero runtime dependencies.

Generates complete AI songs (vocals plus full arrangement) from a theme or from provided lyrics, and instrumental BGM, through a bundled xiapan-music skill and a xiapan_music DSH tool; also queries the Xiapan API balance. Uses a user-supplied Xiapan API key (BYOK), and saves the generated mp3 files on the DSH host.

Hands-free voice for DSH on Windows: local wake word, local Chinese speech-to-text inserted and sent into the composer, and spoken replies that can be interrupted and resumed. Requires a one-time run of the bundled install.ps1 (Python speech dependencies plus microphone pick) and a SenseVoice model already downloaded by the DSH voice input (or placed under models/sensevoice-onnx).

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).

Replaces the summarizing context compactor with a handoff note the model writes itself and a fresh window that keeps only the system prompt, the note, and the current turn.

Edit a sent message in place and resend it as a forked chat, keeping the original: change the text, remove, reorder or add images and files, with session image limits checked before sending (the first message opens a new chat in the same workspace). Also @-references pending images by file name, drag-reorders composer attachments, copies messages together with their images, and adds copy, download and drag-to-composer for images in the conversation.

Registers a QwenWork (千问办公) model provider that reuses the desktop client's local sign-in, runs a loopback OpenAI-compatible bridge on 127.0.0.1, and translates DSH model requests into the ACP-style control-envelope conversation protocol the upstream actually speaks.

Prompt enhancement for the DSH Web GUI composer: pick a style before generating (logic-only, structured, or concise), or supply or clear a custom system prompt in settings; a sparkle button or Shift+Tab rewrites the current draft through the harness model, streams the result into an inline preview (original vs enhanced) with a live thinking timer, fills it back with one click, keeps an undo bar, and offers settings for model route, temperature, length limits, context awareness, undo-bar behavior and shortcuts.

The shared UI foundation for DSH plugins, with a live component preview: design tokens, common controls, workspace browsing and notifications maintained in one place — install it to preview every component in a real host.

DeepSeek Harness plugin: a Sourcetree-style Git workspace — branches, diff, history, fetch/pull/push, AI-written commit messages and AI-assisted conflict resolution, with pinned and usage-ordered repositories.

A transparent always-on-top desktop pet for DSH on Linux (X11 / XWayland), driven by the session state. The pet lives outside the browser as its own window, follows the pointer with a distance and per-display limit, shows a live session-status bubble that lists each session task and state, and auto-starts with your session. Linux only (X11/XWayland); needs Python 3 with PyGObject, GTK3, GdkPixbuf and xdotool. Its sprite sheets come from @michengai/dsh-codex-pet (MichengAI/dsh-codex-pet in this list).

A DSH companion with a macOS native pet and a Windows/Linux Electron companion. Shows the DSH account balance on the pet even when idle, plus color-coded task cards, progress and notices.

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.

JetBrains-style Git workbench for DSH: hunk and line-level staging, side-by-side diff with word-level highlighting, commit graph and branch operations, three-way conflict merge, and commit messages drafted with the host model.

Action Hub for DSH: a standard Agent Skill library whose SKILL.md files can run as model-visible Skills or user-only Commands, with groups, project scopes, import/export and a native Alt+A palette.

Auditable mathematical proving for DeepSeek Harness: a local stdio MCP server with 23 machine-checking tools, a bundled workflow skill, and the /prove and /audit-proof commands. proven, refuted and inconclusive are never conflated, and unproven steps are reported instead of hidden. Needs Python 3.10+ with pip install -r requirements.txt on first run; the 23 tools are unavailable until then.

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.

Adds handoff, context-decision and context-diagnose skills, self-contained context_pressure/context_history tools (peak tokens, compaction stats), per-route realWindows with windowSource evidence, routeKey, handoffAt forced-handoff line, and slash commands.

One prompt row carrying two independently switchable halves. Identity: a fixed self-reference restated at the head, at the tail after every other section, and in the runtime snapshot, with an instruction to treat contradicting claims as stale cache. Discipline: ten rules, each paired with the condition a reviewer checks — external evidence before claiming done, change face equal to review face, correctness from named call points rather than timing, one source of truth, extension through the registered seam, runtime validation only at real boundaries, loud failure instead of silent skips, orthogonal outcomes reported separately, tests through the delivered entry point, and answer shape (conclusion first, length set by the change).

Shows the selected provider's balance below the Harness composer and estimates spend from observed balance drops.

Rewrites the composer draft into a precise prompt with a configurable model and a fully editable rewriting instruction; reads the content behind @file, @folder and @session references for context and keeps the reference chips intact; offers a one-click undo that restores the text you wrote yourself, and places the button in either composer toolbar with an adjustable colour and size.

Interactive HTML or flowchart cards inside the conversation: the tool blocks until the user submits, then returns the clicked action, fields, comments and card state.

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

DSH 动效界面插件：抖动信号场背景、蓝图标题与毛玻璃输入框，可在「设置 → UI 动画」逐项开关。DSH animated-UI plugin: dithered signal-field backdrop, blueprint headline and frosted-glass composer.