Plugins

Real session deletion for DSH: trash by default, plan before delete, audit log, residue receipt

Knowledge-planet (zsxq) knowledge base skill: sync posts and attachments, extract docx/PDF text, guide topic category, summary and quote enrichment, rebuild the index, and search. Requires Python ≥3.10 with the packages in requirements.txt plus your own zsxq account cookie.

Skill that switches the displayed reasoning to Simplified Chinese: it enables the built-in experimental reasoning translation, pins provider bing and target language zh, then verifies and troubleshoots the result. The same repository also ships an optional one-install preset plugin for users who want the same setup without running the steps. Requires the DSH 0.2.1 line with the experimental reasoning-translation bundle and a Web or Desktop GUI; the default provider is bing, so expanded reasoning text is sent to Bing.

Qingyun Jianxian (青云剑仙, Azure Cloud Sword Immortal) — a resident bottom-right widget for the DeepSeek Harness Web UI: a sword-cultivator character that reads your account balance and token spend, keeps a local ledger broken down by model, and has 10 switchable sound effects. Reminders cover low balance, a model waiting on your answer, a pending approval, and staying up too late. Ships 23 idle poses plus a 24-frame sword dance, 22 menu-icon skins, six panel tabs, a line-by-line speech-bubble editor, and black or white robes that follow the light or dark theme. Runs entirely locally and registers no agent tools.

ComfyUI studio in a DSH sidebar panel: connects to a local or LAN ComfyUI server, picks saved image workflows, cleans and rewrites prompts, edits LoRAs and nodes, generates images, and previews results with copyable seeds.

Compact card above the sidebar account avatar showing the signed-in DeepSeek account balance, the current peak or off-peak billing window with a countdown to the next switch, and the rate of both windows. The window is computed locally in Beijing time without a network request; the balance refreshes every 60 seconds and on window focus or visibility. zh/en copy, no runtime dependencies.

WeChat channel with a durable serial queue: messages land on disk first, one reply at a time, and unfinished work survives a restart.

Per-session document ledger in the DSH sidebar (Windows only, Chinese-language UI): which documents a session read, created or modified, with time ordering, type grouping, search, two preview modes and a cross-session view.

JetBrains IDE bridge for DeepSeek Harness on Windows: installs as a DSH bundle, publishes the authenticated DSH URL, registers the open IDE project, relays the IDE MCP server, and serves a compact tool-window panel with streaming, approvals, Plan/Act, steer sending, session switching and archived-session filtering. The matching IDEA plugin is attached to the same release; your DSH agent can install it.

Plays a completion chime — the same sound as OpenAI Codex — on turn end or when the agent needs you (approval, question, plan review). Subagent turns are skipped; mute and volume are adjustable.

Off-peak task queue: typing /闲时任务 in a conversation queues a task with no model call at queue time, and DSH runs it in that same conversation during the off-peak hours of the model that conversation uses (DeepSeek off-peak half price, GLM Coding Plan half credits).

Connects a locally installed, signed-in Codex CLI to individual Harness sessions for planning, code review, and optional image analysis; the CLI must be installed and authenticated separately. Verified on Windows; macOS and Linux use the manual CLI install and are not hardware-verified.

Registers one native DeepSeek Harness skill, design-taste-frontend — an MIT port of tasteskill v2 — for designing, redesigning and reviewing landing pages, marketing sites, portfolios, product pages and app screens (dashboards, admin, settings, inboxes). Directions are dealt by script instead of improvised: three of 20 researched art directions, then the chosen direction's hero composition, section blueprints, type system, palette, photography set and motion entry, drawn from a library of 58 hero compositions, 185 section blueprints, 44 type systems, 56 palettes, 19 photography sets and 18 motion entries, remembering recent draws so consecutive projects do not converge. The house style is measured, not asserted — a static lint plus a real-browser inspector at 1440/768/390 checking headline lines, type, colour, rules, shadows, gradients, contrast and overflow — and the agent then iterates on its own screenshots against a juror's rubric, up to three rounds. Openly licensed photos come with title, author and licence recorded for attribution. Node 22+; Chrome/Edge/Chromium only for the visual checks, and without one it falls back to the static lint and says so.

Stock research tools for US, China A-share and Hong Kong stocks via the stockproof MCP server (Python 3.10+; install with pip install "stockproof\[mcp,cn]" — A-share/HK data needs Python 3.11+): news, earnings and announcement reads with cited sources, explanations of price moves, a whole-market A-share screen, a quant lab that backtests ideas on point-in-time history with a locked holdout, a portfolio check-up and a local web dashboard; 23 tools.

Control a local Linux desktop from DeepSeek Harness with 19 tools: screenshot, mouse, keyboard, scroll, window, application, display and state-verification control over a native X11 session with XTEST, plus optional AT-SPI accessibility-tree reads and widget value setting. Not for Wayland. Requires xdotool on the host (XTEST input); screenshots and window control also want scrot, ffmpeg, wmctrl, xrandr and xprop, and the AT-SPI tools need python3-gi, at-spi2-core and libatk-adaptor. Missing packages can be installed from the plugin's dependency panel with sudo.

Local-first bookkeeping plus a separate wealth-holdings ledger: text entries and receipt OCR through an optional local provider, a running balance, a strictly degrading daily profit chain, and a two-section month-end report where cash and wealth totals are never merged.

Quota takeover for DeepSeek Harness on top of Our Free Model: compiles your platform order and pinned model tiers into one ordered fallback chain, skipping platforms that lack the model, sit in cooldown or are out of quota, writes the result back as Our Free Model combos, and after a quota or rate-limit failure re-plans and injects a follow-up continue message into the interrupted session so the turn resumes on the next platform; the chain ends with your configured aggregate fallback model or the official DeepSeek quota.

Delegates mechanical text work (translate, summarize, extract, classify, rewrite) to an Ollama-native or OpenAI-compatible endpoint by file path, so the source text never enters the main model's context. Registers local_transform, local_batch and local_health; per-workspace local-AI levels (none/light/default/high/max) add a local_worker subagent and a plan-only mode; an opt-in SlaveAgentTeams switch discovers mounted model routes, runs parallel child agents with node-level failover, and registers the endpoint chain as llm-pi-ai routes. Configured from the Web GUI; requires a reachable endpoint.

Floating card above the composer with the conversation's spend, the account balance and the current peak/off-peak billing period, plus a per-answer cost figure under every finished turn.

Read-only GitHub profile page inside DSH — repo cards with language and stars, plus a one-year contribution heatmap.

Lets the conversation AI maintain a plain-language functional requirements checklist in a dedicated session view. A non-official rebuild of Songran241/dsh-req-workbench that keeps the upstream package name; do not install it alongside the upstream original.

Requirement-to-delivery workbench: paste or upload a requirement, and a plan-dispatch-verify loop delivers it with child agents and test evidence.

Declarative DAG runtime for DSH tool plans with bounded parallel execution, read-only retries, per-step checkpoints, cancellation, and conservative crash recovery.

Four filter chips on their own row at the foot of the sidebar, each carrying its match count beside the label: running sessions, sessions waiting for your approval, plan review or answer, sessions with something unread (a turn that finished unseen, or newer activity the panel was not opened on), and sessions updated since local midnight. A chip opens a popover listing its matches grouped by workspace, and a row opens the session. The row ends with a DeepSeek service status dot (green operational, yellow partial degradation or maintenance, red open incident, grey when the feed is unreachable) that subscribes to the status Atom feed once a minute with an ETag conditional request and shows the current finding, the refresh time and the three newest incidents on hover; clicking it opens status.deepseek.com. A separate panel adds an all-sessions chip, a more menu for the last day, three days, week or thirty days, and a manage mode that searches sessions by title or path, multi-selects them, lists the Markdown digests it wrote to the session-digests folder of the DSH data directory in a resizable, full-screenable right-hand column that previews them in place, and either moves the selected session directories to sessions/.trash or deletes them permanently after writing the digest and asking for a second, explicit confirmation. The session you have open cannot be selected. Every panel row also carries a star (shown on hover, filled once starred) and the panel adds a favourites tab that lists only starred sessions, so a long list can be triaged without adding another chip to the sidebar foot.