Plugins
Browse, filter, and install DeepSeek-Harness plugins.
321 plugins found
CHENGXIAO
mypanda-hash/chengxiao
Pairs two DeepSeek Harness machines with a one-time code - directly on a LAN, or across networks through an end-to-end-encrypted relay with a zero-config official default - so one agent can run tasks and exchange files on the other while each machine keeps its own files in place.
dsh-word-vault
wangzhanchao883/dsh-word-vault
English word bank for one or more children, built for DSH 0.2 and newer (0.2.0-rc.1 and 0.2.0-rc.2 verified, including the desktop shell; 0.1.7 through 0.2.x all work). 1.10 added pronunciation: in the memory cards and in the online quiz every word and every example sentence has a speaker button that reads it aloud with the built-in Windows voice - offline, with no network call and no model call - and the whole batch of audio is pre-rendered in the background whenever cards or an exam are generated, so a click starts instantly. 1.11 adds: pick words in bulk straight from the library (10 at a time, top-down, up to 50 selected, with a live counter); turn the words a child got wrong in a quiz into a one-click review set - an online card page that speaks plus a printable PDF; and every card example sentence now carries its Chinese translation underneath. Words marked on textbook or workbook photos (highlighter, red pen) are found by their marks, then translated and stored in a local SQLite bank with an occurrence count; the clipboard picker, the photo route and dictation all land in the same bank. High-frequency words surface by count, three correct answers in a row flag a word as learned, and any selection prints as a memory card set (a word split into chunks, real IPA per chunk, a grade-7 example sentence, a mnemonic hook, an A4 sheet of eight) in HTML, PDF or Word, with a print-and-answer exam page whose four options are set against each other for meaning. On 0.1.5 the settings panel does not appear - install dsh-word-vault@1.0.2 instead.
dsh-mobile-upgrade
langyo/dsh-mobile-upgrade
Requires DSH 0.1.5-rc.1. Mobile UI fixes for the dsh web profile: narrow-screen drawer and settings tabs, a full-width model menu, a stuck-request network chip, and a narrow-header collector that folds the lineage, subagent and background-job chips into one tappable sheet.
dsh-github-router
angeloszou/dsh-github-router
Read-only GitHub access for agents, wrapping several tools. Inside the tool it automatically probes and selects whichever GitHub access method works locally, and tries network proxies on its own. That cuts down the network and parsing problems an agent may hit when accessing GitHub, and the overhead of it repeatedly trying different approaches.
dsh-Kimi-WebBridge
microherox/dsh-kimi-webbridge
Drive the user's real browser through the local Kimi WebBridge daemon: navigate, find-tab, snapshot, click, fill, evaluate, CDP, screenshot, network, upload, PDF export and tab-management tools that reuse existing login sessions.
dsh-plugin-weaknet-adaptor
jiay98528-dev/dsh-plugin-weaknet-adaptor
Weak-network adaptor: transparent long-backoff model-stream retries (10-min cap), local response replay cache, heartbeat auto-reconnect, and degraded-mode token economy, with a bilingual (zh/en) settings page.
tmem
kentzjh-cmd/tmem
Project-scoped deterministic execution trace memory and zero-token session digest plugin for DSH (v0.3.0), compiling session histories via temporal causal slicing and AST symbol attribution into guarded L0/L1/L2 playbooks with inline session_id lineage. Registers /tmem (status, list, forget, gc, telemetry, bench, evaluate, inspect, replay, diff, sync), recalls memories in agent/pre-step only on explicit file-path or error anchors (skipping unchanged repeated steps), and intercepts tools/post-execute errors and code edits with same-origin L2 fixes, failure-loop warnings, statistical L1 co-change partner hints, and runtime adoption telemetry (skipping idle read calls). Requires Python 3 on PATH (or TMEM_PYTHON) and spawns python3 -m tmem.cli asynchronously with a 5-second timeout; stores all data strictly inside each project's .tmem/ directory (no ~/.tmem/ global store) and makes no network requests. Needs Python 3 on PATH (or TMEM_PYTHON); it runs python3 -m tmem on every agent step and tool result.
dsh-session-grep
chenshanhan0711/dsh-session-grep
Literal substring search across every conversation in the workspace, from a search box in the conversation header. The built-in index tokenizes with unicode61 and only matches whole tokens; this plugin scans the local session logs directly, so mid-token fragments and two-character Chinese words match too. Searches user and assistant message text (tool output is opt-in), rotates results across sessions so one long conversation cannot fill the list, and opens the matched conversation scrolled to the hit. Read-only over local session logs, no network access.
dsh-locator
perrylink/dsh-locator
Whole-disk file-name search for DSH with a self-built cross-platform index, agent tools and a Web panel; no Everything required. It indexes your home directory by default; the first content search also builds a local full-text index (SQLite in the platform cache directory) of text and source files there (by extension, up to 1 MB each, 200,000 files) and of your recent DSH session messages; locator_export writes a SQLite copy of the index to any absolute path it is given, replacing an existing file when overwrite is true. The engine is a prebuilt Go binary shipped for six platforms (source in engine/), and the plugin makes no network requests.
dsh-video-studio-skill
znlgis/dsh-video-studio-skill
Registers the video-studio skill in the DSH skill catalog. It builds a Chinese-narrated video from a script: edge-tts synthesizes the narration, the timeline is derived from each line's measured audio duration rather than estimated from text length, Pillow renders every frame deterministically and pipes the raw frames to ffmpeg as rawvideo, burned-in subtitles and SRT/ASS export come from that same timeline, physical-model sound effects and a generated score are mixed sample by sample with numpy, and QA runs in four layers: container, signal, temporal and page level (the page layer only when the browser engine is used). An optional Playwright-driven browser engine renders the whole film as Canvas2D instead, sending PNG frames through image2pipe. Needs network access for the narration, ffmpeg on PATH and Python with numpy, Pillow and edge-tts; the browser engine also needs Playwright with a local Edge/Chromium.
dsh-go-plugin-adb
jminn0/dsh-go-plugin-adb
Gives a DSH agent a real adb shell on the phone itself over loopback, with no root and no PC. Needs DSH Go with android-tools-adb installed in its Ubuntu environment and a one-time Wireless debugging pairing; adb has to be switched to TCP port 5555, which then also accepts connections from the local network (still key-authenticated).
dsh-textbook-parser
teanuky/dsh-textbook-parser
Breaks a STEM textbook into a four-level card map (book, chapter, section, knowledge point) in a sidebar panel, where each point names the defect of the previous approach it fixes and each chapter keeps its remaining gap as the motive for the next; marks key points with a star, shows theorems as cards with statement, conditions and symbol table, folds step-by-step derivations, and exports the same map as a self-contained offline HTML file, Word or PDF. Ships a textbook-parser skill and a textbook_archive tool (save_book / save_chapter / list_books / get_book / delete_book) plus a bundled 12-chapter Signals and Systems sample with 116 sections, 263 points and 601 worked examples. No runtime dependencies and no network access.
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.
xiaoman-memory
yunkong-134/xiaoman-memory
Plain-text memory for DSH where every event is one Markdown section with six fields, recall works by element association rather than vector search, at most two old titles surface per turn, and nothing calls a model or the network. Writing the library stays manual: you can open it, grep it, correct it and carry it to another machine.
dsh-browser-agent
hermes-ai-2026/dsh-browser-agent
19 browser_* tools that let agents drive DeepSeek Harness's built-in sidebar browser over a loopback bridge: accessibility-tree reads with stable refs, real CDP clicks and keys, screenshots, network interception, device emulation, a cookie vault, document-start script injection, multiple browser instances, and raw CDP passthrough. Requires DSH launched with --remote-debugging-port=9222.
dsh-progress-caption
wanted09527/dsh-progress-caption
A one-line caption above the composer that tells you what the AI is doing right now: the step number, the action (e.g. running npm test), and whether it is still going, done, or failed with a short plain-language reason. Inferred locally from tool calls and the thinking text; no model calls, no network requests, no writes to DSH data.
dsh-poker-night
dingchenhui0618-arch/dsh-poker-night
A no-limit Texas holdem table where you are the only human. Eight seats: you plus seven LLM agents written on AI-founder personas (Altman, Dario Amodei, Musk, Liang Wenfeng, Jensen Huang, Lei Jun, Zhang Yiming). Each agent bets through the currently selected host model and falls back to its own heuristic style when no model is available; every agent only ever sees its own hole cards, because the prompt is built from that seat alone. A hand ends by revealing the winner hole cards, with a flip animation replayed on every community card, and a rail keeps a live action-and-table-talk timeline. LAN mode (off by default, single-machine play is unaffected) lets another install on the same network join this very table: the host shows a six-digit room code, UDP broadcast discovers nearby tables, and a joiner randomly takes over one bot seat, with the seat handed back to its character on leave or after twenty seconds of silence. The room code is never put on the wire and every snapshot is generated for one viewer seat, so nobody, not even the host, can read another player hole cards. Fork of BubblePtr/dsh-holdem (MIT), credited in the README and in the UI.
dsh-zspace
439436269-ctrl/dsh-zspace
Turns a ZSpace (极空间) NAS into native agent tools: list, stat, search, read, download, upload, write, create folders, rename, move, copy and delete files on the NAS. Works across networks through the desktop client relay — no LAN, no WebDAV, no SSH — and switches to the faster direct WebDAV transport automatically when the NAS is on the same network. Deletes require an explicit confirm: true; write, upload, create-folder, rename, move and copy run directly, and zspace_write overwrites a same-named file unless overwrite is false.
dsh-windows2macos
twinsearth/dsh-windows2macos
Run one instruction on every online Windows and macOS machine, across networks and regions: a dependency-free relay talks to each machine over outbound connections only (Tailscale, a public VPS behind TLS, or a tunnel), every machine runs the command in its own checkout, and results come back with a commit, a working-tree fingerprint and a command hash so an environment mismatch is reported as unverifiable instead of being mistaken for a different result. Sending work needs its own operator token, separate from the per-machine device token.
bkn-dsh
openbkn-ai/bkn-dsh
OpenBKN business knowledge-network context for DeepSeek Harness
ModelTester
yuer6327/modeltester
Model attribution panel for the dsh web chat — local, zero-network structural fingerprinting (template token leaks, glitch canaries, tokenizer usage-vector matching) that scores which of 29 AI vendors a session's responses most resemble, with a per-claim evidence ledger and a 12-probe active battery.
dsh-mol (dsh-bundle)
yijian-quiet/dsh-mol
Registers the dsh-mol stdio MCP server, built on RDKit with no network access, into a DSH profile. It exposes 12 tools: three-level SMILES validation with human-readable diagnostics and character positions, molecular properties, drug-likeness and structural alerts, structure rendering to PNG/SVG, multi-molecule grids, format conversion (SMILES, InChI, InChIKey, MOL, SDF, formula), standardisation with a per-step change log, deduplication, SMARTS substructure matching, fingerprint similarity, and batch clean-up that reports every rejected row. The patch only registers the server; the tools live in this repository's Python package and must be installed first — the README gives a `pip install` route and a from-source route, and until the package is on PyPI the from-source one is what works.
dsh-netassist
edge-echo/dsh-netassist
Network and proxy doctor for DSH: checks system proxy settings, probes local proxy ports, detects TUN adapters, tests GitHub, TCP and HTTP reachability, scans hosts conflicts and prints concrete suggestions for what to change.
dsh-conversation-archive
yxv1203-collab/dsh-conversation-archive
Manage archived DSH sessions on Windows with search, batch restore and deletion, AI-assisted file retention, and verified local or network-directory backups.