Plugins

RSS/news reader in the dsh sidebar: built-in source library + custom feeds (real-fetch validated on save) + OPML import; the fetch pipeline auto-translates foreign items and LLM-scores new items, generating a daily digest and hourly briefs to cut through information overload; with data backup/restore and preference customization.

Pre-production studio for AI short dramas: imports a TXT/Markdown/DOCX script or pasted text, splits it into episodes, structures the bible, characters, scenes, props, episodes, shots and prompts, compiles prompts for the selected video-model dialect (Seedance 2.0/2.5, MiniMax H3, Kling, Wan), and hard-validates IDs, beat durations, dialogue capacity and look/identity locks before generation, and warns when a shot looks like it has more than one action.

Adds a search-engine picker, a bookmarks panel with multi-level folders (add/edit/delete/import Netscape bookmark HTML) and keyword-then-Enter search in the address bar of the sidebar Browser tab. Requires dsh-better-sidebar >=0.20.0 (the client injects its betterSidebar service and will not load without it). Bookmarks are stored in a local JSON file; no external browser required.

Database workbench inside DSH: import the connections DBeaver already stores (data-sources.json, decrypting saved passwords locally) or register SQLite, PostgreSQL, MySQL and Oracle sources by hand, then hand them to the agent through eight mcp__db__ tools — list_datasources, list_tables, describe_table, sample_rows, execute_sql, explain_query, get_ddl and test_connection. Sources start read-only; one global switch enables INSERT/UPDATE/DDL, every call is row-capped and appended to a JSONL audit log. Also covers groups, bulk actions, a read-only schema browser and a bundled Python MCP server whose dependencies install from the panel. No DSH files to edit.

Import Agent Skills from Git repositories, group them by repository, and enable or disable each one in DSH settings.

Left-sidebar PPT workbench: imports an existing PPTX element by element (theme colours, gradients with alpha, rounded shapes, master/layout decorations, chart pages rasterized) and renders it as real 16:9 slides, lets you edit text and tables directly on the page, and hosts an AI chat column that rewrites the same deck.json.

Local LoRA and QLoRA fine-tuning in a panel: import JSONL, see a VRAM estimate before training starts, evaluate, and export an adapter. One-click Python venv setup, but you must supply your own Python 3.10+ and an NVIDIA GPU (6 GB or more). Verified end to end on Windows 11 with an RTX 3060 6 GB; the installer refuses non-Windows hosts, so Linux, macOS and other GPUs are unsupported for now.

Renders Excalidraw architecture scenes inside DSH — import a local .excalidraw/.json file or an inline document with no account, or refresh from Excalidraw+ over MCP (refresh needs an Excalidraw+ MCP bearer token in mcpToken or EXCALIDRAW_MCP_TOKEN); pan, zoom, select and export SVG/PNG.

Self-narrative memory in five budgeted segments (800 chars, not a full context): a first-person note the model writes to itself, a profile of you aggregated from approved facts, environment facts kept separate from who you are, heat-ranked conclusions, and habits. Every injected fact is traceable to a fact id and can be retired; the profile can be edited by hand, and self-notes stay plain lines you can clear. Facts fade by use and time without ever changing their credibility. Preference-like lines you say are auto-detected as candidates (never auto-approved), with conflicts kept side by side. Recall over past conversations with citations, an incremental index that only decodes newly appended frames, a pointer-required warm layer, a doctor that says which part is broken, a heartbeat proving injection is alive, and export/import so the whole thing moves to another machine.

Loads a workspace's VSCode/Copilot AI configuration into every DSH session: `.github/copilot-instructions.md` is injected as standing instructions, `.github/instructions/**/*.instructions.md` are gated by their frontmatter `applyTo` and join once the session has touched a matching file, and `.github/skills/*/SKILL.md` are registered as DSH skills. Configuration directories outside the workspace, by default `~/.copilot`, are read as well, and a settings page edits that path list.

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.

Windows XP Luna skin for the DeepSeek Harness Web UI: a blue taskbar sidebar, Luna title bars, translucent conversation and plugin windows with adjustable opacity and blur, and a settings page for importing a wallpaper. Light mode only: the skin removes itself when the host switches to dark.

Neuron-inspired memory for dsh: turns buffer into a hippocampus, a nightly patrol classifies fragments into pools, and each turn injects a minimal auto-compacted handoff, the live mood (via dsh-feiyu-mood), and only relevance-filtered memories (Ebbinghaus half-life decay; recalled memories are reinforced like spaced repetition). Hard caps bound storage; past-session original text is searchable over multi-frame zstd; dsh-fish-memory stores can be imported. Optional vector retrieval (off by default, keyword fallback). The 3D panel offers four themes and pulses a signal from the center to each recalled memory.

Import a public DeepSeek or Doubao chat share link into a DSH workspace as a static, resumable session.

An offline novel reader: imports .txt and .epub into the right sidebar, shows the chapter being read in a resizable peek window above the composer, and resumes each book at the paragraph and in-paragraph position it was closed at.

Xray-core supervisor and router for DSH with VLESS import, global/China-direct/whitelist routing, a three-card live panel, and self-update.

Multi-role group chat panel for DeepSeek Harness: join as one character, chat with 2-3 AI members, route replies by addressee with probabilistic bystanders, use @ handoffs, and keep private memory/relationships, a global character library, failure diagnostics, retry of failed replies, message search, a sticky scene note, paged long-history rendering, cross-tab sync, provider-catalog models, and JSON import/export.

Action Hub for DSH: a standard Agent Skill library whose SKILL.md files can run as model-visible Skills or user-only Commands, with groups, project scopes, import/export and a native Alt+A palette.

Local knowledge base and document search for DeepSeek Harness with full-text search, PDF/Word/PowerPoint/Excel/Markdown import, OCR, ACL, and a Web panel.

MCP installer settings page for DeepSeek Harness: paste or import an MCP configuration, review safety risks, then write the servers into the current profile's cordis.patch.yml.

Local-first character archive in the DSH sidebar: it interviews you about someone and writes the answers into Markdown cards, imports SillyTavern character cards, writes a novel chapter by chapter, and roleplays with a card plus its world book; anything written into a card has to be quotable from material you wrote.

Novel writing panel for the DSH sidebar that binds a local manuscript folder, imports Agent Skills as analysis lenses, and scans platform rankings into story-structure outlines.

Give DSH AI timestamped video frames, audio and lyric analysis for Bilibili, YouTube and local files, with encrypted browser-session import, continued viewing, saved site resolvers and native cache management. Requires FFmpeg and ffprobe on PATH or configured in the plugin settings.

Native Anki workspace backed by one real headless Anki collection: browse, edit, add and review notes, suspend cards, import and export APKG, and sync with AnkiWeb, with the same collection exposed to the anki AI tool. No Anki desktop app, AnkiConnect or Qt is started. Supports Windows x64 and macOS 12 or later (Apple Silicon and Intel). The matching headless Anki core, CPython and all Python runtime dependencies are bundled, so no separate Anki or Python installation is required.