Plugins
Browse, filter, and install DeepSeek-Harness plugins.
246 plugins found
dsh-open-in-app-fix
yanli-nyan/dsh-open-in-app-fix
Removes the inherited ELECTRON_RUN_AS_NODE variable from the environment the open-in-app package hands to launched editors, so the desktop host's "Open in ..." launches stop failing with HTTP 502 launch-failed; app.asar is untouched.
dsh-desktop-sidecar
111xjyuan/dsh-desktop-sidecar
Native Windows desktop control for DSH, 21 tools: UIA control tree plus offline OCR to see the screen, click a control by its name, type CJK without touching your clipboard, and drive WeChat / Office / browsers / online-class pages. It never steals your mouse or keyboard (it waits for you to stop), never resizes your windows, lets you type corrections mid-task from a sidebar panel, and after every action it reads the screen back and tells you honestly when it could not confirm. Persistent built-in browser channel. No driver, no extra runtime.
dsh-xmlans
xmlans/dsh-xmlans
Desktop pet for DSH: 兰酱, an angel-winged maid in a transparent always-on-top window who reacts to session states (thinking, working, waiting, done, error), shows task progress and DeepSeek balance, and responds to drags, head pats, pokes and wing touches.
HwFee-dsh (system-fonts)
hwfee/hwfee-dsh
Separate font slots for content, code and terminal text, plus an interface font, an interface-size stepper and an inline-code size stepper, with line-height and letter-spacing controls. Typography only — it never touches colour.
dsh-prompt-enhancer (dsh-prompt-enhancer)
jianyiag/dsh-prompt-enhancer
Composer prompt enhancement for DeepSeek Harness: a button beside the model selector rewrites the draft through one auxiliary LLM call and writes the result back in place, with composer undo left intact; a /prompt-enhance slash command returns the rewritten prompt as a command result line without touching the draft; the enhancement call can be pinned to its own fast provider and model in the profile patch; the rewrite policy is faithful - original intent, language and information are kept, code, commands and paths are preserved verbatim, and nothing is invented or padded with placeholder text.
touchstone-dsh
gavinc-cn/touchstone-dsh
AI-agent-driven automated testing platform for DeepSeek Harness (dsh): drives the harness agent through long-lived multi-round sessions to write and run Markdown test cases, file bug reports, fix the code and re-verify the fix, with a per-project serial queue, a development board and built-in load testing.
dsh-mobile-css
ook826092-cloud/dsh-mobile-css
Mobile adaptation for the DSH Web UI: compact typography, full-width composer with caret fix, drawer sidebar with overlay-close, two-page settings flow, responsive tables and stats, safe touch targets.
dsh-plugin-multimodal
shinjiyu/dsh-plugin-multimodal
Advertise image paste on text-only DeepSeek routes, describe attachments with a vision sidecar, and leave native vision models untouched.
dsh-plugins (dsh-mobile-ui)
tzhr-invest/dsh-plugins
Mobile UI for the Web GUI on narrow screens: full-width responsive layout, overlay session drawer, 44px touch targets, safe-area support, and reading enhancements with zero desktop impact.
dsh-observation-journal
cavan-ou/dsh-observation-journal
Zero-touch runtime telemetry for DSH: every session auto-writes task, model tier, tools, failures, duration, status into a human-readable journal with a stats section (pure observer — no tools, no LLM calls, no injection).
dsh-space-approve
lingyun3241/dsh-space-approve
Answers a pending DeepSeek Harness permission prompt with Space (allow once) or Escape (reject) without focusing the card first, and leaves typing untouched.
dsh-remedy (plugin)
jjjjjjjjnnjnn/dsh-remedy
A patch set for DeepSeek Harness that repairs fused tool-call argument text at the llm/stream seam and re-throws whatever it cannot prove: a call is dispatched only when every fragment parses to a plain object, the first call keeps its id and synthesized ones are marked, and a stream that cannot be split keeps its original error instead of a success. It also listens on agent/request-error to name the deterministic refusals it will not retry, answers per failure class which patch covers it and whether that patch is switched on in the profile you point it at, and reads multi-frame zstd session logs offline to count every turn ending and every error code. Everything it claims is registered with the seam it touches and the limits it does not cross; the resolver never invents a value and never downgrades a configured setting.
dsh-app (dsh-mobile-tunnel)
qiaohelenjoe/dsh-app
The mobile client for DeepSeek Harness: use the DSH running on your computer from your phone. The Android app signs in with the same account, picks this computer from the device list and opens its DeepSeek Harness UI on mobile. The desktop plugin handles registration and the connection, and nothing on the computer needs an inbound port. The UI adapts on its own: on a phone it becomes a touch-friendly layout, and returns to the desktop layout when a foldable opens or the window gets bigger. A PC-side panel shows login state, device quota, device list and the app download QR code.
dsh-eni (armor-break)
konshinhaoshin/dsh-eni
Per-model jailbreak armor: a Settings table arms model patterns with spells (25 spells: 20 across 11 model families, plus a model-agnostic armor, an upstream preferences file and three stackable persona layers), previews the exact injected block, and writes it into the armed model's system prompt only, leaving every other model untouched.
dsh-novel-beat
drizzlegig/dsh-novel-beat
Let the story go where you decide. Beat puts the plot of your novel back in your hands: whose fate turns, which relationship shifts, when a secret comes out, which major turn the story takes. Decide the road, revisit every decision the book has made, and see what the alternatives were. Want to change a turn? The Beat panel lays out every decision by chapter, previews exactly what a change touches (prose, established facts, later chapters), then reverts safely and lets you choose again. Past routes stay readable, so changing your mind no longer means losing the old draft. Off by default; it never writes your prose.
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-adapt-packs
lugabeeee/dsh-adapt-packs
Adaptation packs for third-party DSH plugin panels on phones — 44px touch floor, width reclaim, shell hit targets and floating-widget clearance — loaded by the dsh-mobile plugin from its extensions directory (dsh-mobile must be installed for them to take effect), with a settings page that installs them and a per-plugin compatibility list.
dsh-video-skills
1368546410qq/dsh-video-skills
Browses the awesome-claude-video-skills catalog (373 entries) inside DSH settings and installs a chosen skill into ~/.claude/skills and/or ~/.dsh/skills: it downloads the GitHub archive of that repository and copies the skill directory, and executes no code from the repository. Installs run as background jobs with live progress (bytes downloaded, files copied) and can be cancelled; skills installed by the plugin can be uninstalled, while directories without its marker are left untouched. Rows show an approximate repository size reported by GitHub. The same actions are exposed to the agent as the video_skill_search, video_skill_install and video_skill_uninstall tools.
dsh-bot-toggle
jackwoods-code/dsh-bot-toggle
A floating Bot / ADE switch for the DSH web UI. One click moves between ds-bot's Bot surface and the classic DSH Agent surface, while ds-bot keeps running with its data untouched.
dsh-load-agent-rules
acus9345/dsh-load-agent-rules
DeepSeek Harness plugin that loads the rule files other coding agents use (Cursor .mdc rules, Copilot instructions, Windsurf, Cline, Kiro, Roo, GEMINI.md) with their glob scoping: always-on rules up front, glob rules when matching files are touched, the r
dsh-browser-use
ricardochen1996/dsh-browser-use
Browser provider for DeepSeek Harness: hand it a web task — open a page, click, type, fill a form, read what is on it — and it drives a real browser step by step, seeing an indexed table of what can be done on the current page instead of raw HTML or screenshots. Each step is one bundled-Jev decision rather than a full model turn (1.1 s median per step versus 2.8–3.6 s for chat-model turns in the repo benchmark), and the main conversation receives only the conclusion, so its context does not grow with the page. Targets come from the page it just read and a changed page is re-read rather than clicked blind, and it runs in the browser that Session owns, leaving your everyday Chrome untouched unless you switch to attach mode.
dsh-conversation
huangfuren/dsh-conversation
Conversation outline panel for the DSH web surface: user questions as root nodes (index, time, tool count and text volume per turn) with the Markdown headings of assistant replies hanging beneath them, plus level slider, search scope, bookmarks, copy and export as Markdown, section jump, keyboard navigation, a right-edge density rail and per-session reading-position memory. Read-only UI: it never touches prompts, tool results or provider traffic, and writes nothing outside its own localStorage keys.
dsh-voice-alert
loyalchiiina/dsh-voice-alert
Voice or sound cue when a turn finishes and a failure cue when a tool call or turn errors; user aborts stay silent and subagent sessions are skipped by default. Three modes: voice (Volcengine TTS with cloned or preset voices, fingerprint reuse so unchanged text is not re-synthesised), effects (20 built-in sounds assignable to done, fail and approval, with per-row preview and no API key; the approval slot has no automatic trigger, so it plays only manually) and off. Plays through winmm without touching system volume. Windows only. Adapts DSH kernel 0.2.2-rc.1.
dsh-reel
lsjspl/dsh-reel
Local media library management and playback powered by dsh Web: responsive on PC and mobile with touch swipe reels, instant seeking and on-the-fly transcoding.