Plugins

Run one deterministic skill-routing pass on every user instruction, so the skill catalog is consulted before the model answers instead of after it remembers. On the host it scores the live catalog from the instruction's keywords and the catalog metadata, with no LLM or embedding call, and appends a single producer-tagged message to the step: a ranked shortlist by default, the winning SKILL.md body once the score clears a threshold, or an instruction to call a named router skill exactly once. Unrelated instructions stay silent.

Collapses the DSH Web sidebar panel menu (plugins, extensions, scheduled tasks, IM) from a vertical labelled list into a horizontal row of 36×36 icon buttons with hover tooltips; a Settings → General option switches back to the list and is remembered per browser.

Cone crusher selection and crushing-plant design for MACKORN NH/NS single-cylinder hydraulic cone crushers: requirement-form intake, equipment selection, plant configuration, proposal generation, product-size simulation and field-data calibration. 21 tools, 6 skills, an MCP server, no runtime dependencies.

Draggable floating ball that shows the current DSH session todo checklist live — done/total on its face (capsule mode also shows the running task title) plus a matching progress ring. Click to expand or fold the full list, right-click to switch capsule mode, and state colours mark in-progress (pulsing), done, pending and empty. Replaced plans are archived as history rows that can be hidden one by one, cleared and restored instead of counting towards progress, and other sessions can be pinned side by side. Data comes from the runtime sessions projection first with the official todo panel as a fallback, so the full list survives even when the official strip is collapsed. Text colour can be forced to auto, black or white, the ball snaps back to the bottom-right corner on demand, and it stays draggable right up to the top edge of the window. 130 shared skins — including two official dark and light skins that follow the client light/dark theme automatically, plus a per-ball sync switch so each ball either follows the shared skin or keeps its own — with Shadow DOM style isolation and position remembered per profile, identical on desktop and web.

Built-in Obsidian-style knowledge base for DSH: agents distill conversations, documents and web sources into a structured .wiki graph, with retrieval, graph export, note editing, incremental builds and multi-vault management.

Full-screen lightbox for chat images: click any image to open the full-resolution original with zoom (wheel, buttons, shortcuts), drag panning, prev/next paging, filename caption and download, plus keyboard-arrow and touch-swipe navigation. One "image display control" popover next to the composer collapses every image in the conversation into a thumbnail grid whose column count (1-12) is configurable and remembered, or hides them entirely so the space closes up and restores them later; images are styled in place and no node is ever moved or reparented. The host-side display_image tool takes a local path or directory (TIFF/HEIC transcoded, originals still saved on download), a remote URL, or a keyword search over Bing, Huaban and Sogou that samples loadability before picking a source, re-saves images blocked by hotlink protection and drops dead links after a magic-byte check, returning paste-ready Markdown. It also keeps a system-prompt rule that every delivered image must appear in the reply body as  markdown, because collapsed tool rows are invisible to the reader.

Requirement-driven dsh plugin discovery and environment recipes: plugin_guide matches natural-language needs to a curated plugin directory with reasons and install commands; recipe installs whole setups (通知全家桶, 安全审计套装...) from bundled community bundles with ordered plans.

A green interrupt button beside the send box: silently stops a running agent, which then summarizes its progress and asks for your new requirements, with a customizable interrupt prompt.

Incremental file mounting with read dedupe: mounted line ranges are never re-sent to the model, on-disk changes invalidate and remount, with a Mounted Files tab and token-savings accounting.

Project-scoped cross-session memory: PROJECT.md snapshot injected into the first user message, a memory_remember tool, and auto-reflection after ReAct tasks; each project keeps its own memory file.

Turns one journal paper PDF into a Chinese group-meeting literature report deck (.pptx) in the Nanyang Normal University physics template, plus a matching Word speaker script. Requires Python 3 with python-pptx, python-docx, Pillow and pymupdf (pypdf fallback); install them with pip install -r requirements.txt.

Browse 90 prompt methods with usage guidance and sources, insert a selected requirement, or compare and edit a full rewrite before applying it in the DSH composer.

Self-narrative memory in five budgeted segments (800 chars, not a full context): a first-person note the model writes to itself, a profile of you aggregated from approved facts, environment facts kept separate from who you are, heat-ranked conclusions, and habits. Every injected fact is traceable to a fact id and can be retired; the profile can be edited by hand, and self-notes stay plain lines you can clear. Facts fade by use and time without ever changing their credibility. Preference-like lines you say are auto-detected as candidates (never auto-approved), with conflicts kept side by side. Recall over past conversations with citations, an incremental index that only decodes newly appended frames, a pointer-required warm layer, a doctor that says which part is broken, a heartbeat proving injection is alive, and export/import so the whole thing moves to another machine.

Long-term memory for dsh with defense-in-depth hardening. Two tiers (project `.dsh-memory/` and user `$DSH_HOME/memory/`), each a plain hand-editable JSON file with bounded entries and tags. Writes are atomic (unique temp file + rename, previous revision kept in a bounded history/) and serialized across processes by a per-tier lock with stale reclaim; a disk-revision check refuses lost updates instead of silently overwriting another writer. Memory text is wrapped in explicit untrusted-data boundaries (forged wrappers stripped on write) to blunt prompt injection from remembered content, load-time validation rejects out-of-bounds stores, and symlink escapes out of the tier directory are refused with MEMORY_STORE_ESCAPE. No content-heuristic filtering and no confirmation step, so headless auto-consolidation stays intact. Install: dsh plugin --profile web add github:RocoFish/deepseek-memory.

Adds a chevron row under New Session that folds the sidebar panel list in place, returning the height to the workspace and session list. The hiding is driven by one attribute on the page root plus CSS: no host bundle is patched and no row is unregistered, and plugin entries injected outside the panel slot fold too because they are matched by their own data attributes. The state is remembered per browser and the control hides itself in the 56px rail.

Four-tier prompt enhancement for the dsh web GUI: base / standard / professional / system — the system tier always emits a LangGPT-structured Markdown prompt, the iterate workbench rewrites an editable base under an optional requirement, and literal ancho

A Web3D energy orb decoration for the DSH Web GUI: a particle core wrapped in three mechanical rings that shrinks and falls silent while the session is idle and pulses in a double-beat heartbeat while a turn runs. UnrealBloom glow, draggable with its corner, position and size remembered, and three.js served locally by the plugin instead of a CDN.

Bundled Pixiv skill backed by a Python CLI: search illustrations, manga and novels, browse daily/weekly/monthly/rookie/original/AI/R-18 rankings, look up works, novel series and artists, and download originals, with multi-page works merged into PDF, ugoira converted to GIF and novel series merged into TXT; requires Python 3.9+ with requests, img2pdf and Pillow installed (pip install -r requirements.txt), and R-18 content needs your own Pixiv cookie.

Scientific figures as code: agents create a versioned figure IR, patch it incrementally (one node at a time — untouched objects never move), and export publication-ready, true-vector SVG / PDF / PPTX. Seven papercanvas_* tools (create, get, patch, layout, lint, render, inspect) over a deterministic engine with FC001-FC008 quality lint and CJK-aware font resolution. Verified on DSH 0.1.5-rc.3 through 0.2.2-alpha.2. By the author of dsh-sessionflow. Requires the PaperCanvas engine installed separately: Python 3.11+ and the papercanvas CLI (not yet on PyPI — install it from this repository, put it on PATH or set papercanvasBin); the plugin is a thin adapter and ships no engine.

Middle-click closes a session or a sidebar tab in the DeepSeek Harness Web UI. A session closes through the row's own archive action, so the confirmation for running work and the undo notice stay exactly as the Harness renders them; a tab closes through its own close control, leaving the rule that a lone guide tab cannot be closed untouched. The gesture settles on press-and-release on the same element, so a middle drag never closes anything. Two switches on the plugin's own Settings tab turn each half off, remembered per browser.

Windows only: a night-and-moon themed floating button (a sleepy DeepSeek-chan whale mascot, awake/sleeping artwork) in the bottom-right corner arms the plugin — click it to power the machine off once every root Agent session has gone idle, and click it again to disarm. The button is draggable (position remembered in localStorage), automatically dodges other plugins' corner popups, and survives the frontend re-rendering it away. When the batch finishes a 60 s cancellable countdown starts (configurable 10–600 s) with the remaining seconds on the button and a matching banner: clicking Cancel runs shutdown /a and disarms the mode, and any new session starting during the countdown aborts it. A right-click menu on the button switches the shutdown condition between only-on-success and shutdown-regardless, persisted across restarts in $DSH_HOME/deepseek-chan-sleep-shutdown.json. Zero runtime dependencies.

Remembers the reasoning effort you explicitly chose for each provider and model and restores it when you switch back, without replacing DSH's own model picker.

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.

Per-conversation backup for DSH: incremental session mirroring, fuzzy restore (preview then confirm), and rule-based cleanup with a 7-day recycle bin.