Plugins

A drawer panel for the DSH web UI with a paginated list of free LLMs; one-click or batch writes each as a custom model provider into Settings, plus a local multi-key rotation proxy.

Literary DSH plugin: one session is one book. Cover-card /new opens one of 19 realms, a prose-only 诸天万界DSH preset keeps the model writing fiction, keyword worldbook entries ground each turn, and /export-story typesets the session into a Markdown novel.

A live preview panel for the ArtCraft applications inside the DSH Web GUI: the agent step list, the current render, a filmstrip of earlier steps and a timeline for time-based crafts, fed by the craft plugins' renders.

一个可以让ai自动调用VLM(多模态模型)进行视觉分析的dsh插件。A dsh plugin that allows AI to automatically invoke VLMs (multimodal models) for visual analysis.

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.

A grey ASCII ocean behind the whole window: whales swim continuously across the sidebar and the chat area, so a creature can drift from one column into the other. Once a conversation starts, every creature leaves through its own edge — up, down, right, or left — while the sea calms, and the sidebar keeps one frosted surface whose height follows its content, so a passing whale becomes a soft shadow behind the labels; an empty conversation brings them back in from the edges. Streaming replies reveal from blur to clear, and reduced motion falls back to a still frame and the native renderer. Neutral ink only, with day and night palettes.

Add PaddleOCR vision to text-only LLMs like Agnes, DeepSeek, MiniMax, GLM and Mimo. Supports local files and URLs. Free tier of 20,000 pages/day; needs a PaddleOCR API token you apply for (free, 1-3 business days).

Synesthesia Encoder for DSH: a vision model translates images into compact structured spatial text (canvas/elements/percentage coordinates), giving text-only LLMs pixel-level image understanding via the `mm_vision` tool.

XMUM Moodle in the sidebar: courses grouped by term, a schedule merged from the calendar, assignment and quiz deadlines and dated announcements, and batch downloads filed by term, course and material type. It signs in to the school Moodle (l.xmu.edu.my by default) with your campus ID and password, keeps only the resulting Moodle token in the DSH credential store, and its HTTP routes have no DSH access check.

Adds a left-sidebar panel to the DSH Web GUI: one ranked board of the films and series released in the past 21 days or due within the next 31 days, taken from JustWatch for the US, five European countries, Japan and Korea and from Douban for China. Each row carries a poster (or a drawn placeholder when the poster CDN will not load), a synopsis of five sentences or fewer, Douban and Rotten Tomatoes ratings and popular Douban comments. A search box filters the board as you type and asks Douban, TVMaze and the iTunes storefront about a title, merging the three into one record and listing the matches already on the board separately. Synopses, comments and titles are translated through the host translator service, so the Chinese/English switch re-renders the board. No API key and no account.

Audits and evolves agent context at four scopes: a single repository, a group of related repos, a knowledge hub, or the whole machine. Discovers projects, instruction files, skills, tools, knowledge, state, evidence and credential references; reports drift, duplicate facts, broken pointers, budget overruns, unindexed knowledge, stale state and unregistered credentials; scores each scope on its own dimensions; and turns findings into reviewable migration plans that back up every file before moving it. The format contract is a versioned spec (CTX rules, markdownlint-style IDs) following open conventions — the AGENTS.md open format, llms.txt-shaped indexes, YAML front-matter note headers — now backed by published JSON Schema contracts (draft 2020-12) for the hub manifest and the plan JSON, plus an editor modeline that validates hub.yaml with zero configuration. The hub is declarative (schema 2, zones with light profiles, sections may live outside the hub root), a one-pass check command runs scan, doctor and score, and state writes are atomic and lock-protected; scoring stays honest about missing evidence — a waiver greens the gate, never the score. Ships a stdlib-only Python CLI plus two skills; credentials are referenced by path and never printed. Requires Python 3.9+ (stdlib only) and DSH 0.2.1-rc.2 or newer, below 0.3.0.

MiniCut tools for making multi-shot AI short films through the MiniCut CLI: status, next step, and run, with a price shown and the user's yes required before any paid job.

Adds a right-column panel that compares a set of images the user picks from the workspace — 1/2/3/4/6 per page, per-image zoom and pan, the zoom and viewed position of one image synced onto the whole set, and a reorderable filmstrip. The browser loads every image, so none of it reaches the model.

DSH 实时库文档：写代码前自动拉取项目安装版本的官方文档，消灭幻觉 API。零配置零 Key。Version-pinned live library docs for DeepSeek Harness — kill hallucinated APIs

Rate-limit plugin for DeepSeek Harness: FIFO queue per provider/model bucket with minIntervalMs and maxConcurrent

Generate llms.txt for dsh plugin repositories from package.json + README — AI-readable discovery files, zero runtime deps, read-only by default. CLI plus an agent-callable llms_forge tool.

LLMTR AI gateway provider for the DeepSeek Harness LLM seam: 170+ chat models and Türkiye-hosted LLMs behind one OpenAI-compatible key

Image-to-text input for the Web UI: paste or drag an image and it is transcribed into structured text and sent, giving text-only LLMs image-input takeover (OpenAI-compatible vision API).

Vision for text-only LLMs: image description / OCR / VQA via free Gemini and GLM vision APIs.