Plugins
Browse, filter, and install DeepSeek-Harness plugins.
839 plugins found
dsh-doubao
ralsei1337/dsh-doubao
Rewrites DSH model output as "Doubao-speak": a barrage of superlatives, a long list of negations, healing diagnoses, verbatim repeats and a service-style closing question. Five strength levels (off / light / standard / heavy / max) instead of a dial, two independent channels (system-prompt style bible and llm/stream post-processor), three transform modes, and every length knob adjustable live in a settings page with a preview.
dsh-skill-panel
ztqing4agents/dsh-skill-panel
Native-style skill manager for DSH web GUI: manage built-in, local, and market skills in one list with lossless toggles, local/GitHub import, and a built-in SkillHub marketplace. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-design-library
ztr66/dsh-design-library
Design and video library for DeepSeek Harness: 716 indexed entries covering 256 design systems, 97 visual styles, 124 shot recipe cards and 204 craft documents, plus computed review, render-environment checks and overlap/overflow detection.
dsh-inline-diff (plugin)
him666233/dsh-inline-diff
Inline diff cards for every edit, write and str_replace_editor call: old/new line numbers, character-level highlighting, 57-language detection with a red/green-safe syntax palette, a side-by-side viewer, Read-row line ranges, and a settings page for display style and highlighting.
dsh-local-trim
leonlee114514/dsh-local-trim
One-step local model onboarding plus a light harness for DSH. On load it probes loopback and private-network inference endpoints (llama.cpp, Ollama, vLLM, LM Studio), reads each model's context window and output budget from the backend, and registers the provider through settings, so a local model becomes selectable without hand-writing config. It then trims only that provider: masks the tool list through tools.restrict, drops named system-prompt sections and memory contexts, swaps the injected instruction block for a short one while preserving its first 80 and last 96 characters, and truncates tool-schema descriptions without touching structure or required fields. Cloud providers are returned as the same assembly object, untouched. The injected block is re-read from disk on every assembly, so editing the file takes effect on the next request with no reload. The panel in the plugin detail page is a live operations view: it shows whether the takeover is active and by which hook, how many characters the last assembly was trimmed from and to, and which providers are online (probed concurrently, with a real one-shot completion reporting latency and reply text per provider). If a local backend is not running, the panel offers a one-click launch: it probes the port first (never starting a duplicate), then starts the matching server (ollama serve, or a custom command you supply) detached and polls until the endpoint answers, clamping the wait to 3-600 seconds, reaping the child it just spawned if it never answers, and naming the paths it searched when it cannot find the binary. A stop button kills backends this plugin started (SIGTERM, then SIGKILL) and refuses to touch a server somebody else launched. When a model has just been downloaded and no backend exists yet, the panel can also scan the machine for .gguf files with hard caps (300 models, 20000 directory entries, Downloads only one level deep), start llama-server on the chosen one (skipping the .mmproj. vision projection and reusing a server that already answers), register the provider under the model id the backend reports, and run one real completion to prove it works. Before spawning anything it does a fit check that counts only the part of the file the GPU cannot hold: hostBytes = fileSize - VRAM, compared against max(total RAM - 2GB, free RAM). A file far larger than RAM is refused outright; a heavy host-side remainder is refused unless you confirm; the same remainder under mmap is only a warning, because the weights land in reclaimable page cache. A load-mode selector (auto / mmap / none / mlock / mmap+mlock) maps to llama.cpp --load-mode, and the default is whatever the engine picks. From 0.4.2 everything above is derived from the platform instead of assumed: home and data directories come from os.homedir() and the OS conventions (LOCALAPPDATA on Windows, ~/Library/Application Support on macOS, XDG_DATA_HOME elsewhere); backend binaries are looked up in the official install locations of each OS and can be pinned with OLLAMA_BIN, LLAMA_SERVER_BIN, VLLM_BIN or LMSTUDIO_BIN; model discovery walks DSH_LOCAL_MODELS_DIR, ~/models, ~/llama.cpp/models, ~/.cache/llama.cpp and the LM Studio data directory, with the WorkBuddy layout demoted to one candidate among them. VRAM is asked from nvidia-smi, then rocm-smi, then sysctl hw.memsize on Apple Silicon (flagged unified, so the same bytes are not charged twice), and any flip to a unified-memory budget is handled explicitly. Before building a command line the plugin runs llama-server --help once and keeps only the flags that build actually has, so older engines get --n-gpu-layers 99 and no KV quantization while newer ones get auto plus q8_0; anything it cannot probe is passed through rather than dropped. LM Studio is started with lms server start --port N rather than being lumped in with the engines that need a model argument, and vLLM gets its own --model example instead of llama.cpp syntax. Windows, macOS and Linux are all exercised in CI, including a windows-latest and a macos-latest job, because a cross-platform claim that has only ever run on one OS is not a claim. Two of those suites exist specifically to catch this machine's blind spots: one replays the whole suite with process.platform rewritten to each OS, and one deletes every advantage this development box has — no nvidia-smi, no rocm-smi, no sysctl, an empty PATH, a home directory with no models — and requires graceful degradation rather than a crash. That last one caught a real class of bug: several assertions hardcoded 'lms.exe', which cannot pass where the executable has no suffix. Measured on the reference machine: a 16.9GB 27.2B MoE with 14.5GB of VRAM holds a 285-691MB working set and sustains 76 tok/s after a 1000-2000 token warm-up, so file size alone is not a fit signal. From 0.4.3 the scanned models are shown as a table rather than a dropdown (one row per gguf, with its size and any fit warning), so you can see what is on disk before choosing. From 0.4.4 the delete button on the provider table removes the provider entry, not the weights file on disk. That distinction matters: what people actually want to clear out is the dead entry - one pointing at a port that never answered, one registered twice for the same port, one whose model file was deleted long ago - not a 15.8GB gguf. Deleting a config is undone by importing again; deleting a gguf is not. The removal goes through settings.mutate with an unset op on \['providers', id], because update merges layers: writing null there leaves a null provider key that the loader reads as a broken config rather than an absent entry, and only mutate/unset is implemented as Reflect.deleteProperty. The result is read back afterwards, because unset on a key that comes from a lower layer (the profile cordis.patch.yml or a home patch) does not delete it - it restores the inherited value, so the call succeeds while the entry stays. When the entry is still there the panel says so and points at the file to edit instead of reporting success, and a DSH whose settings service predates mutate() is told plainly rather than silently ignored. Deleting takes two clicks and clearing the confirmation is as easy as clicking another row; on success the online and baseURL caches are cleared and the tool surface is recomputed, since the provider list changed. The gguf itself is never touched by this plugin - the scanned-model table is display and selection only.VRAM reporting is driver-dependent. From 0.4.5 model discovery also reads Ollama's own storage: its blobs are named sha256-… with no extension, so extension matching could never list them. Weights are now accepted by extension OR by GGUF magic header (only dot-free files of at least 8MB, capped at 600 header reads per scan; obviously-not-weights extensions like .json/.md/.safetensors/.part are never opened). Ollama models are resolved from their manifests: the model layer is picked by mediaType application/vnd.ollama.image.model (never adapter/mmproj layers), the human-readable name is restored from the manifest path (…/qwen3/8b becomes qwen3:8b, …/latest stays bare), tags sharing one blob merge into a single row, and the metadata kv supplies name, quantization and context length. Adopting an Ollama model does not start a duplicate llama-server over the same weights: it confirms ollama serve answers on 11434 (launching it if not) and registers the provider under the tag the endpoint itself reports — the blob's sha256 name would never match anything, since 11434 lists qwen3:14b, not sha256-a8cc…. Two real-machine-only traps fixed on the way: existsSync('<file>/') returns true on Windows, so manifest-vs-directory branching now goes by parseable content instead of trailing-slash semantics; and the quantization label table had been copied from a stale ggml enum generation, labelling ggml_type 15 as Q5_K_S while ollama show reports Q4_K_M. Same-blob dedup is by (dev,ino), keeping whichever entry has a name. Verified on the reference machine: the scan goes from 1 row to 4 — qwen3:14b (8.64G) and qwen3:8b (4.87G) now appear with Q4_K_M badges — in 4ms; npm run verify exits 0 (launcher 57, portable 35, panel-host green including the linux/darwin simulated-platform runs); the new panel-host suite builds a real miniature Ollama store in tmpdir and points OLLAMA_MODELS at it, so results no longer depend on whether the test machine happens to have Ollama installed. From 0.4.6 the panel shows each model its real name instead of its file name, because some download tooling stores every weight under one generic file name, leaving the real name only inside the GGUF header. The scan now reads each file GGUF KV section - not a fixed-size head buffer, because tokenizer.ggml.tokens (hundreds of thousands of strings on a large MoE model) and the chat template push general.file_type past any window; and GGUF string arrays are interleaved (u64 length plus payload per element), so treating them as a length table walks the cursor off the rails and produces astronomically large reads. Reading and skipping are accounted separately: key names and types are always read, string values over 256KB and string arrays are only walked past without copying, non-string arrays over 64 elements are skipped outright, and a cursor past the file size stops the read, because an out-of-range position means the cursor is wrong rather than that a limit should be raised. The displayed name goes Ollama tag, then general.basename plus size_label, then general.name, then the file name, with the full string kept for the hover text since it usually carries internal build numbers; mmproj files are now identified by general.type as well as by name. A short read (truncated file) no longer feeds half a value into the parser. On the reference machine, reading one file header parses 44 keys in about 16ms, and a whole scan takes about 31ms. From 0.4.7 the smoke test no longer fails the first time and passes the second. The old code gave both request attempts one shared AbortController and one shared 30-second timer, so the total budget covered both rounds: a cold load of a large model burned it in round one, the abort left the signal permanently aborted, and the second round (the Ollama-style reasoning_effort: none retry) could not even be sent - which is exactly why the first click returned This operation was aborted and the second succeeded. Now each attempt gets its own controller and timer, a timeout retries once automatically (retryOnTimeout, on by default) because a first-request timeout almost always means the weights are still being read from disk, and non-timeout errors (refused connection, HTTP error) are not retried at all. Timeouts report themselves: the message says the wait exceeded smokeTimeout, whether a retry already happened, and points at the new smokeTimeout setting (seconds, default 60, clamped to 5-600) - port responding is not the same as tokens coming out, so it is a separate knob from launchTimeout. Unrelated: an example naming one specific model was dropped from the panel copy, since not every machine has that model downloaded, and the explanation now describes the mechanism instead. From 0.4.8 adopting a plain .gguf no longer registers an unrelated Ollama model. The port field used to follow the last click rather than the file: selecting an Ollama entry moved it to 11434, selecting a plain .gguf did not move it back, so a 11434 left over from an earlier click stayed. The server accepted it as given, probed the port, found Ollama answering and borrowed it (the skipped-launch log line about the port already answering is this path), and then the model id - derived from the file name - could not be matched against Ollama's /v1/models, where the fallback was to grab the first entry in the list. Embedding models tend to sort first, so the adopt registered a model the user never selected. Now the port follows the file: a plain .gguf landing on another engine's port (ollama 11434, vLLM 8000, LM Studio 1234) is moved to 8090 and the panel says so. Only ports belonging to a known engine are overridden - a hand-typed 9999 aimed at your own llama-server is left alone. The model-id fallback no longer guesses either: it takes the only entry when there is exactly one (a single-model backend may legitimately report a different id than the file name) and otherwise reports an error listing the ids it did see, rather than registering a wrong provider. From 0.4.9 the panel can no longer vanish. It used to be possible for the whole section to disappear from the plugin detail page while the host stayed up, the HTTP bridge kept answering, and no crash log appeared. The cause was a scope error: a helper was declared inside one component and referenced from another, which is valid JavaScript and passes a syntax check, so it only threw at render time - when the user pressed the scan button and React ran the state updater. The host wraps every slot entry in an error boundary whose fallback is an empty anchor, so the throw was contained silently: the panel was not uninstalled or crashed, it was replaced by an empty div. The helpers now live in the module scope both components share. The client half also gained its own test - a minimal React stand-in loads the real panel module, really calls the components, really runs the state updater at render time, and presses the scan button, so this class of error can no longer reach a release.
dsh-obsidian-notes
revive123456/dsh-obsidian-notes
Obsidian-style notes panel in the DSH right sidebar: a tree over a local folder, Markdown preview with line numbers, edit and save, Obsidian callouts, tables and wikilinks, screenshot paste into the vault, one-click insert of a note path into the chat input, and batch archive or delete for sessions and workspaces.
dsh-qq-sounds
gexiubinhelloworld/dsh-qq-sounds
QQ-style alert sounds for DSH: the QQ message tone when a turn finishes and the QQ system tone when approval or an answer is needed, reading your local QQ install on Windows and falling back to built-in tones.
companion-memory
root-7-khx/companion-memory
Roleplay companion memory for DeepSeek Harness: per-conversation editable persona documents (1-100 weight per entry, AI may refine them in small story-driven steps), per-speaker favorability and bond in multi-person chats, active memories with AI-scored vectors, a nine-emotion table updated every turn, an AI self-evolution ledger, recent 48-hour transcript search, companion-style conversation compression, sleep-time compute that pre-guesses the next topics while a conversation is idle, per-conversation tone cards learned from its own past replies, a habit-hint small-model bridge that writes the user habits it extracts into this conversation memory, a date-archived local knowledge base with auto-packing by turns or tokens, and a unified integrations page that lists curated plugins and downloads their GitHub releases. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-nonead-prompt-recall
noneadchina/dsh-nonead-prompt-recall
Shell-style prompt history in the DSH composer: ↑ in an empty composer recalls the previous prompt and keeps walking back, ↓ walks forward to the empty draft; history is kept per workspace, persisted across restarts and records only human input, while typing or editing a draft returns ↑/↓ to normal caret movement. Adds a Settings page to export, import and clear the history.
dsh-pet-chat
lintwosix-beep/dsh-pet-chat
Desktop pet with a chat panel: QQ-style history, favorites, and branch threads that merge back into the log as a forward card; the panel docks beside the pet and resizes from its edges and corners. Modified from PC2005-cloud/dsh-pet (MIT); do not install together with the original.
dsh-mobile-shell
htao-uu/dsh-mobile-shell
DeepSeek App-style mobile shell for the DSH Web UI with drawer navigation, app-like composer and settings; best paired with the dsh-android app, which supplies the native Android back key (standalone use works but is not recommended).
dsh-skin
changan0617/dsh-skin
Homepage-style chat background skin for DeepSeek Harness, ported line by line from deepseek.com/harness: WebGL fluid background (5-colour noise field, flow-field brush, light, bloom, vignette, grain), white 90px dot-line grid with mouse perturbation and spring return, and a digital point-cloud whale that assembles from particles, breathes, swims, is pushed radially by the cursor and brightened by a light beam. Dark theme only; disabled automatically under reduced-motion.
dsh-cherry-bridge
saxem1997/dsh-cherry-bridge
Bridge DeepSeek Harness sessions to Cherry Studio desktops over the LAN - QR pairing, 15 cherry_* tools for agents, sessions, history and interactions (including uploading local files as attachments and reading desktop-uploaded images back as image blocks), a settings page for devices and per-workspace tool scope, and an experimental subagent that delegates the way DSH's own subagent does (Claude Code / Codex style) with the child running on a Cherry Studio agent.
dsh-token-total
gaohu1997/dsh-token-total
Token usage dashboard and per-session tree statistics: provides time-range filtering, 6 summary metric cards, GitHub-style activity heatmap, model stacked-bar charts with cache hit rate curves, and Top-5 model/provider rankings; enhances the native session status bar to display aggregated token usage across main agents and all subagents with detailed breakdowns.
dsh-session-groups
xyzyaemiko/dsh-session-groups
Edge-style tab groups for the DSH sidebar on Web and Desktop, with nested groups up to three levels, workspace tree, single-list view, recursive sorting, RGB colors, collapsed status summaries, subagent expansion with status, and an approval-gated agent tool for organizing groups.
dsh-preset-doctor
drscrewdriver/dsh-preset-doctor
Preset doctor for agent presets: scans home/profile `cordis.patch.yml` declarations and legacy `.agent-presets/` dirs across host lines 0.1.0→0.2.0, shows every detected preset as an official-style card (source/era badges, broken red border, per-finding details, current DSH version + era catalog), and syncs a preset to the current syntax per-card behind an explicit confirm dialog (.bak backup + atomic write; report-level findings are never written — startup scans are report-only). npm: `dsh-preset-doctor`.
dsh-selection-quote
xt765/dsh-selection-quote
Cursor and GitHub PR Review style selection toolbar: quote selected text to chat (Cmd+L), write inline comments (Shift+Cmd+L), and batch-insert from a composer dock.
dsh-prompt-history
wenaixi/dsh-prompt-history
Claude Code-style prompt history, double-Esc picker, selection quote/copy & right-click paste.
dsh-claude-slider
yicun0316/dsh-claude-slider
Replaces the reasoning-effort dropdown with a draggable snapping slider, and draws 13 canvas effects across five style families (thrust, fluid, fantasy, tech, texture) with three tiers that unlock as the effort level rises. Custom mode combines 10 particle forms with 8 flow trajectories; 9 accent presets and a hex picker recolor every effect, and optional key sounds cover two voice clips, a synthesized duck squeak, a mechanical click and a local audio file.
DSH-Desktop (whaleharbor-promo)
moonlitdropofblood/dsh-desktop
Turns a plain-browser DSH session into the WhaleHarbor lite client: a standalone browser --app window with a client-style top bar and task notifications, while the full desktop installer downloads in the background (multi-source, SHA-256 verified) and an install prompt appears when ready.
dsh-concise
hoyyang/dsh-concise
One-click Concise output style for DeepSeek Harness: results first, no filler, same work depth. Every final reply opens with a plain-language digest card (select text to copy); file paths inside the card are highlighted and click-to-open with the OS default app, and a deliverables zone lists the files the turn actually produced from real tool outputs. Scoped per session, works in headless/CLI too, with /concise command, style.md overrides and a local API.
word-docx-writer
remplexi/word-docx-writer
Generate and format Chinese Word (.docx) documents from a simple markup file, using built-in heading styles, SimSun and Times New Roman fonts, and per-chapter figure and table captions.
deepseek-style-theme
lichtspur/deepseek-style-theme
DeepSeek homepage-style theme skin for the dsh web GUI: fluid particle background, glass sidebar and composer, running-subagent progress panel, and DSTT peak-hour color modes.
dsh-paste-code-block
cjm-m/dsh-paste-code-block
Cherry Studio-style paste: multi-line/indented/fenced text or code pasted into the DSH Web composer collapses into a bordered, collapsible, language-tagged card and is restored as a fenced code block on send.