COMMUNITY REGISTRY
Inspect versions, compatibility, HMR behavior, source code, and exact install commands.
dsh-markdown-previewIn-chat preview for produced files in DeepSeek Harness Web: click a produced-file chip to render Markdown (server-side markdown-it + highlight.js), images, or plain text right in the conversation; system-app open and reveal-in-folder stay one click away.
dsh-web-icon-indicatorBrowser tab favicon reflects the DSH session state: idle / running / asking / done, recolored and animated from one base SVG in the browser (fully configurable colors + effects). · 浏览器标签页 favicon 实时反映 DSH 会话状态:待机 / 运行中 / 提问 / 完成,由单个 base.svg 在浏览器中动态上色与动画(
dsh-connectAll-in-one DeepSeek Harness (DSH) chat bridge: one plugin, one config block — Feishu/Lark, Telegram, DingTalk, and Web mirror adapters plus the web-settings pane, all behind a channels[] selector
dsh-arknightsUnofficial non-commercial Pramanix and Eyjafjalla astral garden skin for DeepSeek Harness Web
@aa2246740/dsh-skillhubDSH manager for user Skills already on disk in Agent home and DSH home
@max-null/dsh-draft-polishDSH draft polishing: Click ✨ to the left of the Send button to call the LLM and polish the draft in the input box. Optionally include conversation context to help the model understand the intent. Uses the DSH llm channel and can also be used while a conversation is in progress.
dsh-openrouter-liveOpenRouter provider for DeepSeek Harness with a live model list: every model OpenRouter serves, with context sizes, reasoning levels and image input read from its /models endpoint.
dsh-cohubDeepSeek Harness CoHub: Chinese agent orchestration — users can choose between 2 work modes (co-orchestrator pure orchestration / cohub-cordis creative mode) + 13 specialized agent skills + multi-model consensus; additionally, the two supporting presets cohub-subagent-readonly / cohub-subagent-write
dsh-enforce-package-managerDeepSeek Harness plugin that catches the wrong package manager for the repo (npm install in a pnpm repo, pip install in a uv/poetry project, and so on) and gives the model the exact right command.
dsh-pre-duediligencePre-visit due diligence and company research in DeepSeek Harness with Host-tracked progress, fixed-route Qichacha MCP calls, entity confirmation, and downloadable reports.
dsh-mcp-manager-plusMCP manager settings page for DeepSeek Harness: list, enable, configure and add MCP servers from the browser, written durably to the profile's user patch layer.
dsh-inline-diffRenders edit and write tool calls as always-expanded side-by-side diffs styled like git diff, with optional syntax coloring and word-level highlighting.
dsh-web-search-zaiZ.ai and Zhipu web search for DeepSeek Harness, with Coding Plan MCP quota, REST API billing, and bilingual settings.
dsh-market-gist-autosyncDSH plugin: back up your DSH configuration to GitHub Gist on a schedule. · 把 DSH 配置定时备份到 GitHub Gist。
dsh-usage-miniUSEAGE WINDOW: a tiny DeepSeek Harness usage dashboard. Currently supports Claude and Codex subscriptions, plus DeepSeek API only.
dsh-thinking-token-statPer-turn thinking-token statistics for DeepSeek Harness, folded once by a durable session projection and read by the composer dock.
dsh-agent-pickerAgent mode selector: choose an external Agent CLI (Claude Code / Codex / Qwen / Gemini / Aider / Goose / Hermes / OpenClaw / Ollama…) as the default executor, like selecting a reasoning level. The Agent catalog is data-driven: adding a new Agent only requires adding one data entry, with no code chan
dsh-quicksightTwo-tier image reading for text-only models in DeepSeek Harness: fast local OCR (RapidOCR, offline) first, then a vision model via modlens as fallback.
@fonlan/dsh-model-routerDSH plugin: a virtual model-router provider that aggregates every configured provider and model, routes each request to the model's currently active provider, and ships a settings page for per-model provider ordering and switching.
@yangzhe1991/dsh-web-enhanceDSH Web UI enhancement plugin: expand the chain of thought by default, track session costs for the official DeepSeek API, monitor GLM Coding Plan quota and fixed-cost safety thresholds, open files with the system default application, completion notifications (sound alert + tab title), and retain out
dsh-opencode-go-plusDSH plugin: standalone OpenCode GO LlmAdapter — owns its provider route outright (per-model wire protocol incl. union-alpha's anthropic-messages), manages models/context/modality/effort declarations in its own settings namespace, never touches llm-pi-ai
dsh-memoriesDeepSeek Harness dual-ledger cross-session memory: auto-extracted long-term facts (MEMORY.md) plus a living project progress ledger (PROGRESS.md), recalled into every new session.
@summersec/dsh-web-authTransport-level authentication gate for the DeepSeek Harness Web GUI
@mars.liu/dsh-rich-editorThird-party dsh web plugin: a rich Markdown notebook for the composer region — tool-row toggle plus a CodeMirror editor card with Codex-style list editing, submitted through the conversation service
@bittersmilezzz/dsh-model-selectorEnhanced model selector for DeepSeek Harness: single-level menu (search + grouping) + bottom inline reasoning intensity (Effort) slider.
babeldoc-on-dshBabelDoc-on-dsh: a DeepSeek Harness plugin bundle that translates PDFs through BabelDOC while keeping their original layout.
dsh-jev-guardPre-execution safety valve for DeepSeek Harness: a TypeSafe Jev question decides whether a shell command/script would irreversibly delete or overwrite real data. Four states (allow/revise/block/escalate), mounted on tools/pre-execute. Zero dependencies; W
dsh-swarmdropSend files from your DeepSeek Harness agent straight to your phone, and reference what your phone sent back — no account, no public IP, end-to-end encrypted.
dsh-data-cleaning-agentData cleaning and data enrichment for CSV/XLSX/JSON enterprise lists in DeepSeek Harness, including spreadsheet cleaning, deduplication, profiling, optional Qichacha MCP and exports.
@ztlovelsw/dsh-model-profileDSH Web GUI model capability configuration: Directly set reasoning efforts for each configured model in the model catalog editor under “Settings → Models,” and sync reasoning efforts and context capacity in batches or according to models.dev presets, writing them to the llm-pi-ai settings namespace.