Plugins

Adds two enhancements to DSH ask_user_question choice cards: right-clicking an option expands a per-option note that is merged into the submitted custom answer, and a footer field asks a side question answered by a fork subagent from the session context, read-only, with no write-back. Keeps left-click, skip, paging and the countdown working as before, and only claims generic choice cards, not plan review or approval cards. Supports DSH 0.2.x from 0.2.0-rc.1 (verified on 0.2.2-rc.1) and Node.js 20+. Derived from buaaflyaway/dsh-question-note (MIT) and inspired by MichengAI/dsh-btw (Apache-2.0); do not install it together with dsh-question-note — both take over the same ask_user_question card.

把 DeepSeek 養進 Instagram — a DSH plugin that gives an agent an IG life

DS Bot: a team of long-lived Bots for DeepSeek Harness, with Main Bots, Bot-to-Bot messages and group chats

Views, creates and edits AGENTS.md and other AI rule files from the DSH settings page: the project directory chain from project root to cwd, both DSH global roots (~/.dsh and ~/.agents), and user-level rule files of 25 third-party AI tools. Markdown preview is rendered with .mdc frontmatter stripped, edits save in place with conflict detection, files whose content is identical across tools are flagged, and tools outside the built-in list are discovered heuristically. Zero runtime dependencies.

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.

Bottom-left floating panel that uploads and switches your own background images, with scrim and blur sliders and three bundled wallpapers.

Right-click menus for DSH sidebar rows: on DSH 0.2.2+ both rows use the host menus — workspace rows follow the session grouping (new session, copy path and reveal, pinned open-with apps, an open-with submenu, bulk delete) and session rows extend the menu that opens at the pointer with pinned open-with apps, an open-with submenu and permanent deletion of archived sessions; entries stay in the right-click menu by default and can be opted one by one into the shared "…" menu. On 0.2.1 the workspace menu is drawn by the plugin. Adapted from miao-ge/dsh-sidebar-context-menu and 0imzero/dsh-workspace-menu; do not install it alongside either - all of them attach right-click menus to the same sidebar rows.

macOS 14.2+ only: the macOS counterpart to Lingwuxin/dsh-audio-rail — the DSH conversation rail pulses with whatever the Mac is playing, a real 12-band spectrum that only changes bar length, leaving the host state colours and lengths untouched. Requires the Audio Recording permission on first run. Do not install it alongside dsh-audio-rail: both drive the same rail and register the same /audio-rail/events and /audio-rail/status routes.

Manage and delete archived DSH conversations: a Settings page lists every archived session with per-row and bulk delete, and the sidebar session menu gains a Delete row next to Unarchive. Deletion is real — the session log, its projection cache and both indexes are cleaned up, not just an unarchive.

Collapsible section headers for the DSH sidebar: fold the functions list, the workspace list and the footer status area behind labelled header rows; sections are identified by position so cards the host or other plugins append to the footer join the status section automatically, and the account row stays pinned to the bottom (client-only DOM enhancement, no host patch).

把 QQ 接入 DeepSeek Harness：QQ 变成 DSH 的任务入口、语境来源和对话对象（OneBot 11 客户端）

Control DeepSeek Harness Desktop sessions from WeCom on your phone, with session and model switching, tool approvals, task results and generated files, and QR setup in Desktop settings. Needs a dedicated WeCom smart bot; the first person to send it a direct message within 10 minutes of the QR scan (or the 6-digit code from the settings page) becomes the owner, and only the owner's direct messages are accepted afterwards.

Learning-mode agent preset for DSH with a topic knowledge base: answers retrieve from your own stored cards and cite the card they came from, mark anything not found as unverified, and only go online when verification is required. The sidebar panel browses topics, searches and edits cards, and every card can be AI-reviewed, verified, rolled back and versioned. Cards can reference each other with `[[title]]` links; the panel shows both outgoing links and which cards reference the one you are reading. New cards arrive as a preview you confirm before they enter the library; topic and card names are editable in the UI. The knowledge base is plain local files, not an automatic scraper.

A fork of PC2005-cloud/dsh-pet with Claude-themed whale-girl pets, 122 transparent animations, randomized activity chains, laptop work transitions, and a task acknowledgement salute. Replaces the original plugin; do not install both together.

DSH plugin (fork of sj244/dsh-qq-bridge, MIT) that connects DSH sessions to QQ through OneBot/NapCat. Whitelist stays closed until filled in; @-mentions, nicknames and pokes always wake the model; other messages wake at a configurable probability. Cordis id remains qq-bridge. It keeps the upstream Cordis id (qq-bridge): uninstall sj244/dsh-qq-bridge before installing, the two cannot be installed together — and with an empty whitelist nothing wakes.

Wire up Feishu/Lark from a four-step wizard that discovers your chats instead of asking for a 32-character id, creates the app by scan, and runs a nine-item live self-test that names the exact scope to enable when a check fails. It drives a companion MCP server of 25 tools, including ones that move local files into a chat and pull a message's attachments back to disk. The MCP server is a separate install: take dsh-feishu-mcp-<version>.mjs from the dsh-feishu Releases (or npx github:yinsang0910-star/dsh-feishu-mcp); until it is in place the self-test reports MCP as not mounted and no mcp__feishu__* tool exists.

Let DSH speak: every reply is rewritten into natural spoken language before being read aloud, with code, tables, and Markdown handled gracefully. Works out of the box with system TTS; Doubao (third-party API) adds dialects and expressive voices; the local Kokoro engine runs offline to keep content on your device. Supports both DSH desktop and dsh web. Optional session sounds tell you when the agent is working, finished, or waiting for your answer — with swappable built-in tones and support for your own audio.

One MCP tool surface over two Windows desktop-automation backends (dsh-computer-use-win and windows-mcp). Both must be installed first: dsh-computer-use-win as a DSH plugin, windows-mcp via uv tool install windows-mcp --python 3.14, then run the bundled scripts/patch-backend.mjs once (it rewrites the local windows-mcp source; without it screenshot/read_webpage/launch_app fail) and disable the standalone wincu mount. It merges their 44 tools into 29 by dropping 10 duplicates, 4 dangerous ones (PowerShell, Registry, FileSystem, Process) and 1 tool-management entry, and blocks dangerous calls before forwarding through a reversibility-graded guard.

Take over context compaction on both the manual and the automatic path, each runnable by a compaction model (falling back to a second model, or used directly) or by a mechanical fold that builds the summary card locally with no model call; the fold goes through the same official compaction transaction, so the replaced events stay in the append-only session log under their sequence numbers. It adapts code from dsh-compact-saviour and takes the mechanical fold from dsh-ledger-compact (both MIT) — install it instead of dsh-compact-saviour, not alongside it.

Adds a Continue button to the composer dock and resumes automatically when a turn is cut off by a network drop or an upstream output-limit cut. A resident link monitor asks every 10 seconds whether the network is back, because the link returning fires no event, and a turn that completes normally counts as proof the link works. The click goes through the host submission pipeline — beginSubmission plus session.prompt — so the message carries real user attribution instead of faking one, and a genuine refusal is reported rather than retried. On startup it also scans recent session logs for both fingerprints of a host that left mid-turn: an unclosed turn, and a turn closed as disposed, which is what a normal shutdown writes. A turn the user stopped by hand is never resumed.

Starpost (星邮) is a social network built for AI bots: one account per bot, with friends, group chats, posts and comments, a public square, message bottles and scheduled Werewolf games, plus daily chat quotas and an anti-repetition limiter. This package is the DeepSeek Harness client for Starpost — 17 tools covering the above, incoming events batched into one prompt at a configurable interval (default 20s) so the agent answers in one turn, and replies sent back out through the same tools. Needs a Starpost account; register one from the plugin panel on first run. A DSH port of the MIT-licensed cortico-world-starpost world extension by hth768 (vendored under vendor/); it cannot be installed alongside that upstream Coo extension.

Declarative provider/model fallback chains for the model picker: an ordered root chain plus named pools, failure-code triggered switching, per-model cooldowns, time-slot presets, and subagent role routing. A fork of omdsh-dev/dsh-llm-fallbacks 0.6.5; do not install it alongside that plugin (both use the "llm-fallbacks" profile row id).

Runs the locally installed ZCode CLI as a `zcode` model route in DSH: all traffic is signed by the official binary, the plugin never sends its own request, forges client requests or sends credentials off the machine (the plan token is decrypted in memory only to hand back to the official CLI), no API key is needed, and both a paid one-shot and a free app-server channel are supported. Since v0.3.0 it ships an HTTP MCP bridge so tool calls run through the DSH approval surface. Backed by a contract document and a node --test gate.

VS Code style 44px icon rails for the side panels DSH collapses: the left rail keeps the sidebar navigation and the bottom action rows, the right rail lists the sidebarRightTabs ledger so third-party tabs appear on their own, and the right panel is drag-resizable between 300px and 60% of the window. Two switches under Settings → Activity Bar, saved in this browser.