Plugins

DeepSeek Harness plugin that aggressively saves tokens without affecting task quality: lossless head/tail truncation of tool results (default), grep-result folding, pressure-triggered dsh-compaction-basic delegation, and an LLM-summary CompactionEngine subclass for the aggressive mode. Saves 97.6% of input tokens on an 8-scenario synthetic suite, with all reductions routed through dsh event flow so model-visible = logged stays intact.

Per-step model router: named presets with ordered image/keyword-group rules route each agent step across every mounted provider and model, adding collaboration flows for image transcription, an image guard, a multi-plan quota panel, and per-step decision observability.

Live2D companion for DSH: a floating Cubism model with state-driven expressions and mouse tracking; config-driven geometry (size/position from the plugin config); official Haru sample bundled (© Live2D Inc.).

Replaces the model button with a Claude Desktop-style model and reasoning-effort switcher, with per-model effort memory, theme-aware styling, smooth slider feel, rounded track, hover highlight, and pixel-dot animation.

Smart subagent routing and evolution: bind roles to any provider/model registered in DSH, including models supplied by Codex and Claude Code provider plugins, with workspace learning.

CC Switch-style token consumption statistics for the dsh web GUI: per-request usage log, real-consumption hero, request-time-bucketed trend chart, per-model peak/off-peak list pricing, per-project and per-model stats, and account balance.

A slim port of Headroom: compresses tool output, cuts up to 20% of context, without affecting the model's context cache or agent performance.

Session-based task board in the sidebar: drag sessions into todo/doing/review/done columns, create sessions with model and reasoning-effort selection, trash with restore, notes, export/import.

Bottom-right quota pill and panel aggregating membership quotas across AI platforms (Kimi Code, DeepSeek, Zhipu Coding Plan, BigModel, Qwen, Scnet, TokenRouter, SupaWriter), auto-syncing API keys from the DSH credentials domain with colored usage bars.

Enhanced ACP server for DeepSeek Harness: block & reasoning streaming, usage telemetry, model/effort switching, permission presets, session resume/archive, per-session MCP servers, and Zed file/terminal forwarding.

Per-workspace project kanban board: nine kanban_* model tools for agent-driven planning plus a browser board UI, with disk persistence, undo, templates, and archiving.

LLM provider that runs the locally installed Claude Code CLI as the model backend, so requests go through an existing Claude subscription instead of a metered API key; native tool calls are bridged over MCP.

Per-model token usage and cost estimation with DeepSeek account balance, shown in a Settings panel tab.

Rewrites the reasoning row: instead of the raw chain of thought it shows first-person working notes, written segment by segment while the reasoning streams in, with a toggle back to the original. Notes are stored per session so history rows keep them after a reload, the note writer runs on a configured provider or a downloadable local llama.cpp model, and the model still reads the full reasoning.

Exposes a running DeepSeek Harness as an MCP server over Streamable HTTP or a bundled stdio bridge, so an external agent can create workspaces and sessions, prompt, fork, rename and switch the model on them, read and write files, run foreground and background shell commands, call the instance's own agent tools by name, and read per-turn and cumulative token usage with a cost receipt when a price plugin is installed.

Telegram sessions, topics, model controls and approvals for DeepSeek Harness.

In the DSH desktop file preview, click an HTML element (with its source line and column) or lasso part of an image for the vision model to identify, then send your change request to the current session; local HTML edits keep a history with before/after previews, line diffs and rollback.

dsh 插件（DeepSeek Harness / cordis bundle）· 模型接入 —— 把 CodeArts、CodeBuddy / WorkBuddy、LobsterAI、Qoder / Qoder CN、TRAE、Cline、Loomy、Raccoon、MiniMax Code、ZCode、Gemini、OfficeAce、Accio Work 十五个账号渠道接进 dsh 的模型选择器，并保留数据看板、运行日志与本地网关（含局域网中继）。上游 2.1.0 的匿名免费车道、EAC 协付车道、

Connects DeepSeek Harness to QQ over OneBot 11/12, Satori, Milky or the official QQ bot API, with a separate session per group or private chat and nine built-in tools: group management, chat history with image understanding, web search, page fetching, reverse image search and image generation. Optional proactive chatter reads recent messages on an interval and decides whether to join in, with the judging and speaking stages able to use different models. Models, proxy and allowlists are configured in the DSH settings page. First run requires an owner (auth.owners): with it empty the plugin drives no sessions at all (deny by default, so installing it does not expose the bot); once set, private chats and groups work immediately, guests only reach the guest tool allowlist while qqbot and qqconfig stay owner-only.

Drive a dsh agent from Telegram: one chat topic per project directory, voice and photo input, ordered per-topic job queue, per-topic goals and task checklists, session forks and branches, live progress message, command menu in en/ru/zh, and control commands for models, plugins, sessions and logs. Search the plugin registry, read a package card before installing, and update every plugin with one tap. Requires a Telegram bot token and your numeric user id in allowedUserIds.

Long-term memory for dsh via a local Hindsight service: each turn is retained with local timestamps and relevant memories (time-aware, reranked) are recalled into context before replies; each workspace binds to one memory bank, several workspaces can share a bank, and banks can have Chinese names; what gets retained is configurable (default: user messages and AI replies); history is imported only when you choose to; includes a three-step setup page with one-click local deployment (uv + Python 3.12), a sidebar status card and a panel to browse and search memories. Needs a SiliconFlow API key for the default embedding and reranker models; only Linux is verified so far.

DSH plugin for generative UI: the model writes an interface instead of describing one. Declarative DSL compiled and executed in a DOM-less Worker sandbox, then drawn by the host with native DOM; a sandboxed-frame escape hatch for anything the component in

Controls the Chrome you are already logged into: the model clicks, types, fills in forms, reads pages, takes screenshots and downloads files in your real tabs, including inside same-origin iframes. Runs on macOS, Windows and Linux; needs Chrome 116+ and Node 18+. Works over native messaging and chrome.debugger instead of a listening port. Ships 12 browser_* tools and a cb CLI, dispatches real CDP input events, draws an on-page cursor that also appears in screenshots, and adds a toolbar panel with live status, the controlled tabs and a pause switch. browser_download saves a link, an element's target or a URL with the site's login cookies (following redirects, keeping the server's file name) to the Downloads folder or the session's working directory; cross-origin iframes are named in the snapshot but not entered. The native host and the MV3 extension travel inside this tarball: on first load the plugin copies them into place and writes the native-messaging registration itself, so there is no clone, no install.sh and no python3 - the one step left is adding the extension to Chrome: install DSH Bridge from the Chrome Web Store (https://chromewebstore.google.com/detail/bippcbbbipppmhjkjomgdnjbooecbimo, no Developer mode), or load the bundled unpacked copy at chrome://extensions (the plugin opens that page and puts the folder on the clipboard for you). Until then the tools fail loudly with those steps. On Windows the registration goes into the registry (for Chrome and Edge) and the host talks over a named pipe. A daily CI run on GitHub's macOS, Windows and Linux machines installs the plugin, loads the extension into a real Chrome and reads a page through the browser_* tools.

Settings page for the REA reverse-engineering tool gateway (rea-mcp-lite) that lives in the same repository: pick which tool bundles load into context, see the tool count and token cost of each bundle, save and recall named presets of your loading setup, hot-swap them without restarting DSH, and check this repository and REA for updates, including classifying the tools a new REA release added. The page is split into sections with sub-menus, filters and pagination. Requires REA (rea-agents) and the gateway registered as the rea MCP server (see repo README).