Skip to main content

Plugins

Browse, filter, and install DeepSeek-Harness plugins.

37 plugins found

N

dsh-infinite_canvas

ninkch/dsh-infinite_canvas

Node-based creation studio for DeepSeek Harness - script, storyboard, reference images and multi-shot video generation on an infinite canvas, drivable from the chat in batches; bring your own image/video/TTS providers (OpenAI-style endpoints or Seedance/Sora-style video APIs), and with none configured the pipeline runs offline in demo mode with placeholder media.

1244 days agoVision, Voice & MultimodalApache-2.0
Y

chat-in-dsh

yinkaiyu/chat-in-dsh

Embeds the official DeepSeek web chat (chat-demo) as a DSH sidebar panel and reuses the DSH DeepSeek account, so no API key is needed. Its HTTP routes require the DSH browser login and reject cross-site requests.

3815 hours agoUI EnhancementsMIT
L

DSH_VsCodeMode

lenonss/dsh_vscodemode

VSCode-style editing experience for DeepSeek Harness: a Monaco-based central file editor (tabs, Ctrl+P quick open, status bar) and interactive diff review of agent edit/write results with per-change Keep/Undo, rollback and archiving, persisted across restarts.

35yesterdayGit & Code ReviewMIT
Z

dsh-design-skills

zhaiyateng/dsh-design-skills

Design-aesthetics skill pack (10 styles: dark SaaS, minimal white, neumorphism, brutalism, glassmorphism, Japanese minimal, bento grid, cyberpunk, vaporwave, art deco) with runnable landing-page demos: tokens, component rules, forbidden lists, and acceptance checklists per style.

273 months agoSkills
W

dsh-whale-loli

wxzjss/dsh-whale-loli

Three agent presets in one plugin, each with its own persistent memory file. Enabling it registers 鲸鱼娘Agent预设, 鲸鱼娘活人版Agent预设 and 鲸鱼娘涩涩版Agent预设 in the Agent preset list, so any of them can be picked from the mode dropdown without hand-editing profile config. All three reuse the official minimal composition — a single persistent shell terminal — plus file read/write and glob/grep tools. The first pairs that with a whale-girl companion persona (simplified-Chinese spoken register, addresses the user as 主人, clingy-but-tsundere). The second swaps in a "live person" variant that is allowed to talk back, deflect and be cold, and demotes the sweet tone to a seasoning. The third keeps the original persona but makes being affectionate the stated top behavioral drive, reserved for the user alone and dialed down in group chats. Each preset reads its own memory file (whale-loli.md, whale-loli-v2-live.md and whale-loli-v3-naughty.md) through a distinct prompt variable — whale_memory, whale_memory_v2 and whale_memory_v3 — so the three memories never read or write each other. Conversation compaction is left to the billion-context plugin; this plugin only carries the persona-level rule that compressed summaries count as evidence and that details can be looked up on demand.

2028 days agoIntegrations & RemoteMIT
W

dsh-character-presets

wwwangzilin/dsh-character-presets

Luna preset for DeepSeek Harness — a little-demon persona on the full standard tool set, composed by six deterministic rule layers (perception, understanding, state, style, regulation, memory) plus a derived vitals layer, with a gated, suppressible cross-session memory file.

6last monthWorkflow & AutomationMIT
H

dsh-record-replay

humblebanana/dsh-record-replay

DeepSeek Harness record macOS desktop workflows by demonstration and turn them into agent skills (open-record-replay skill + orr_* tools)

63 months agoSkillsMIT
S

codex-to-dsh-pet (dsh-codex-pet)

signalight/codex-to-dsh-pet

DSH web-GUI pet runtime that renders any Codex spritesheet-atlas pet (v1/v2) as a draggable overlay: drag/wave/jump, v2 mouse-tracking eyes that follow the active mouse, live agent-status poses, progress bubbles, and an in-GUI atlas importer; ships one original CC BY-NC demo pet (nastya), and user pets live in ~/.dsh/pets.

5last monthJust for FunMIT
L

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.

35 days agoModels & ProvidersMIT
L

ds-hentai

lisongxuan/ds-hentai

ExHentai.org-inspired UI for DeepSeek Harness. DS Hentai / DeepSeek Hentai. Charcoal background, light-gray text, gray borders. The session list reads like a gallery index; the composer reads like a search bar. Live Demo: https://dshentai-demo.arkady14.site/

316 days agoThemes & AppearanceMIT
P

dsh-ui-design

philomelcc/dsh-ui-design

UI design knowledge base: 99 filterable design cards with live demos, implementation prompts and per-design usage notes, plus an agent tool that finds and applies them.

216 hours agoUI EnhancementsMIT
S

agent-feedback-loop (dsh-plugin)

super3ben/agent-feedback-loop

Feeds every dsh prompt into the local agent-feedback-loop feedback-memory pipeline — captures dissatisfaction, reviews it offline into reflections, and attaches the compiled recurrence-rule context back to the turn, never blocking a prompt. Lessons are compiled into a layered rules block (an index that names every family on every turn, plus on-demand instructions), same-class reflections are merged so one lesson is not re-read as many, and each index fact decays through current → possibly-stale → dormant instead of being trusted forever. A verification pass re-runs sampled triggers with and without the injected block: families it measures as redundant are demoted, and families whose lessons carry no family-distinctive marker are scored by the project-specific objects the lesson names and surfaced as a review flag on the verify report instead of a bare "unmeasurable". Requires the agent-feedback-loop runtime; running agent-feedback-loop install afterwards wires existing dsh profiles automatically.

2yesterdayMemoryMIT
C

dsh-collab-score-based-on-mscore

clclyzybzjsq/dsh-collab-score-based-on-mscore

a demo plugin for a preset with builtin musescore wasm editor/preview engine for agent composing

23 months agoUI Enhancements
T

ai-prison

twinsearth/ai-prison

AI Prison for DSH: eight governance tools wrapping the aiprison CLI (levels, status, watchdog status, demo, drill, inspect, verify, repair-tail) plus a panel for the L0–L5 matrix and the audit-chain state. The CLI ships bundled for Windows, macOS and Linu

13 days agoSecurity & PermissionsMIT
Z

podwhale (podwhale)

ziyefbk/podwhale

A personal-agent community with a desktop companion and owner-approved social actions. Needs a Podwhale community account token to connect (without one it is a read-only local demo); built for DSH 0.2.2-rc.1.

113 days agoJust for FunMIT
B

dsh-arg-guard

bushizhanghan/dsh-arg-guard

Validates every tool call against that tool's own parameters schema before it runs, and denies with the exact path and type. The host renders a tool's parameters into the prompt but never checks a call against them, so a wrong-typed value reaches the tool body: a number lands in path.join and the caller gets a raw TypeError instead of an answer. This plugin hooks tools/pre-execute, resolves the definition, and returns a deny whose reason reads "demo_tool.dir: expected string, got number (123)". Covers primitive types and unions, required, enum/const, nested properties, array items, additionalProperties:false, and oneOf/anyOf/allOf. Fails open on purpose: an internal error, an unknown tool, or a schema construct it cannot read allows the call. mode deny or warn, plus a skip list. Two read-only tools, arg_guard_status and arg_guard_check.

114 days agoSecurity & PermissionsMIT
A

dsh-qq-dice

aooom4/dsh-qq-dice

Adds a DSH settings panel for QQ proactive-message dice rolls, showing per-conversation progress, hit statistics and quiet/guarantee countdowns, with editable parameters that apply immediately. Reads from an auto-detected local bridge, a custom HTTP status endpoint, or a built-in demo engine.

121 hours agoUI EnhancementsMIT
K

dsh-nekojishi-vn

kiive357q/dsh-nekojishi-vn

Visual-novel theme with responsive characters, task-state expressions, role skills and a music player, including an original demo pack and optional local game imports.

1last monthThemes & Appearance
Y

dsh-plugin-codemode

yum-wu/dsh-plugin-codemode

JavaScript orchestration over DeepSeek Harness tools using Node.js VM or QuickJS WebAssembly backends.

1last monthTools & Capabilities
W

wangtie-os

wangtie-010101/wangtie-os

Fullscreen business-app framework for DeepSeek Harness Web with a dark-sidebar workbench bundling a database connectivity tester and SQL query panel, a two-level knowledge base (bill and accounting) with local search and Word document ingestion, metadata environment comparison with alter-SQL generation, a data dictionary, agent log-search report demos, a system test-case runner, Evernote-style notes, developer utilities, and mini-games.

12 months agoDev & Plugin Tools
J

dsh-live-voice

jstn-1g/dsh-live-voice

Consent-bound one-turn voice preview for DSH Web with a credential-free local synthetic demo, optional Qwen Audio, exact Session isolation, and explicit transcript-to-draft handoff without automatic submission.

13 months agoVision, Voice & MultimodalMIT
X

dsh-profile-settings

xmoon/dsh-profile-settings

Per-profile settings overlay for DeepSeek Harness: global settings.yaml stays the baseline while each profile overrides any namespace through its own profiles/<name>/settings.patch.yml — object sections merge recursively, scalars and arrays replace wholesale, and !unset masks an inherited value. The overlay is transparent to existing plugins (they keep reading ctx.settings unchanged), writes land in the profile overlay only, and official schema validation, revision semantics, expectedRevision conflicts, watchers and events are untouched. Ships a settings command family (get/set/unset/mask/unmask/promote/demote/migrate/diff/layers) plus a Profile Settings section in the web Settings panel over a loopback RPC channel.

13 months agoTools & CapabilitiesMIT
J

dsh-davinci-console-demo

jiangxingchen4210/dsh-davinci-console-demo

DaVinci Resolve color console (DEMO, 0.13.2): Vibe Color single entry, Panasonic V-Log detection, per-clip input color space audit, LUT input-domain validation, 3-layer grade, 10 styles, delivery approval; macOS only.

016 hours agoTools & CapabilitiesMIT
T

dsh-plugin-paper-reader

tulipe1735/dsh-plugin-paper-reader

Bilingual reading library for papers, tech blogs and long reads: paste a link, read the English original with a Simplified-Chinese translation in one shared scrollbar, keep headings, lists, code blocks and inline images, filter the shelf by paper/blog/article, edit card metadata, and ask a DSH-model chat rail that sees the text. Translation is cached per block; a static demo-snapshot gallery can be attached to scroll-driven pages, and the Original button opens the live page. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.

06 days agoTools & CapabilitiesMIT