DSH Marketplace

dsh-lark

omdsh-dev/dsh-lark

Lark/Feishu bot channel for DeepSeek Harness: each chat drives its own agent, and tool approvals, model questions, and plan reviews return as cards answered by a button or a reply. Switch workspace and model from the chat (`/cd`, `/model`, `/new`), and run several bots that keep separate sessions and can hand turns to each other in one group.

5717TypeScriptBSD-3-ClauseSource

Install

Add dsh-lark to DeepSeek Harness

Install into

via npm

Resolves a published tarball rather than cloning the repository, and installs without any extra setup. Swap `web` for your profile name if you run another one.

via GitHub · npm package

Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.

What happened when we ran it

Installed cleanly when we ran it

Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run todayon DeepSeek Harness 0.2.2-rc.5.

Show it in your README

install verified — dshmarketplace

For maintainers: the badge serves this listing's latest sandbox verdict, so a re-run updates it on its own — and it links readers to the full result here.

Due diligence

Before you install dsh-lark

  • Source of record: omdsh-dev/dsh-lark — present in the community registry that DSH's own plugin market installs from.
  • Licensed under BSD-3-Clause.

Security

What installing dsh-lark runs

Build scripts

pnpm skipped these dependency build scripts, and the plugin registered without them:

  • protobufjs@7.6.6

Under pnpm 11 the install stops and asks you to approve them instead; an approved script runs with your user account's permissions.

Source: our sandbox install on DeepSeek Harness 0.2.2-rc.5, today.

Install hooks

package.json declares no preinstall, install or postinstall script.

Source: package.json on GitHub, read at our last sync. prepare is not recorded; it runs only on a GitHub install, and shows up under build scripts when pnpm blocks it.

README mentions

Neither shell commands nor credentials.

Source: a keyword match on the README — a hint about what the plugin touches, not an analysis of its code.

Not an audit. These are the record of one install and two text scans; nobody has read this plugin's code on your behalf. Plugins run with your agent's permissions.

The AI take

What it is — DeepSeek Harness Lark/Feishu plugin: integrates chats to drive agents with card-based approvals.

Who it is for — Users collaborating multiple DeepSeek Harness agents in one Lark/Feishu group chat benefit from independent sessions and turn handoff. Users who do not use Lark or Feishu as integration channels can skip this plugin because they lack the chat environment.

Watch out — No public server or callback configuration needed. Tested and registered successfully in a new profile. No obvious issues found.

The verdict — I would install it because no public server or callback configuration is required.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.

What dsh-lark does

Separate chat streams pass through agent sessions, approval gates and workspace folders.

dsh-lark is a DeepSeek Harness plugin that connects agents to Lark or Feishu chats through a bot channel. DeepSeek Harness loads the plugin as a channel, where each chat, topic or sender scope can drive a separate agent session. The service uses a WebSocket connection, so it does not require a public server or callback URL. Messages start tasks; execution details, tool calls and results are sent back to the chat, while questions, plan reviews and tool approvals arrive as cards that can be answered with buttons or replies.

Commands such as /status, /ws, /cd, /model, /new and /stop control the active workspace, model and session. Incoming files are written to .dsh-lark/inbox/<timestamp>-<message-hash>/ in the current workspace; outgoing files can be sent back, with approval required in group chats. workspaceRoots, sessionScope, senderAllowlist, groupAllowlist and approvers constrain how the service is used. Additional bot instances keep separate settings, credentials and sessions, and can hand turns to one another in a group.

dsh-lark suits teams that want to operate DeepSeek Harness from Feishu or Lark rather than a terminal, including workflows requiring human approval during execution. It is a poor fit where chat-based access is unnecessary, where voice transcription is required, or where the documented Node.js and DeepSeek Harness version requirements cannot be met. It also duplicates some host commands, which are passed through to the DeepSeek Harness command runtime.

dsh-lark documentation

Configuring dsh-lark

DeepSeek Harness loads dsh-lark into a profile; the README shows dsh plugin --profile web add dsh-lark-channel for sharing a profile with dsh web. Start-up configuration is read once, so changes require a service restart. Relevant settings include:

Setting Behaviour
workspaceRoots Restricts directories selectable from chat.
sessionScope Selects chat, chat-thread or chat-sender session separation.
senderAllowlist, groupAllowlist, approvers Narrows who can use the bot or approve actions.
instance Names an additional bot instance; the first remains unnamed.
botPeers, botHops Restricts bot-to-bot conversations and their consecutive turns; botHops defaults to 6.
receiveFiles, sendFiles Enable file reception and delivery; both default to enabled.
attachImages Enables image attachments when the selected model supports vision.
maxReceiveFileBytes, maxSendFileBytes Set per-file limits, which default to 20 MiB.

Received files remain under .dsh-lark/inbox/ and are not deleted automatically. Group file delivery always requires an approval card.

Commands provided by dsh-lark

/status shows workspace, model and session state; /ws lists workspaces; /cd <name-or-path> changes workspace; /get <path> sends a workspace file; /model, /model use <provider/model> and /model reset select or restore a model; /new starts a session without changing workspace or model; /stop stops the current task; and /help lists commands. DeepSeek Harness also passes host commands such as /plan, /compact and /permission to its command runtime.

Requirements and operation

The service requires Node.js ^22.19.0 || >=24, DeepSeek Harness 0.1.0-rc.6 or newer, and a Feishu or Lark tenant. On macOS and systemd-based Linux, its user service can be managed with status, logs -f, restart and stop. The service reconnects its WebSocket with rate and back-off controls.

Known limits

Voice messages are stored but not transcribed. PDF, XLSX and DOCX files can be downloaded but not previewed in chat. A group can have at most three pending file approvals, and each file is limited to 20 MiB by default.

Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub

Same category

Alternatives to dsh-lark

dsh-im

xmanrui

2.4k

Connect IM bots to DeepSeek Harness via QR codes or bot credentials (9 channels: Feishu, WeChat, DingTalk, WeCom, QQ, Slack, Telegram, Discord, and WhatsApp).

Installed cleanly when we ran it

npm packageJavaScript28d ago

AI review

dsh-im

What it is — Attaches DeepSeek Harness to nine messengers—Feishu, WeChat, DingTalk, WeCom, QQ, Slack, Telegram, Discord, WhatsApp—signing in with a QR code or bot credentials.

Who it is for — A DeepSeek Harness agent is already running on a machine you left behind, and the missing piece is sending it a line from WeChat or Feishu on your phone. If that agent already sits behind your own web UI or internal gateway, this hop is not what you are short on.

Watch out — On a clean profile it installed and the harness registered it; no obvious pit showed up there. Live send and receive on those nine channels was not part of that run.

The verdict — I only add it after I am willing to let this plugin hold the IM login—getting into a profile is not the decision.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
Details

dsh-pocket

shaobeichen

2.3k

Remote phone access to the DSH Web UI: scan a QR code for LAN or public (cloudflared tunnel) access with real-time sync, a mobile-adaptive layout, and a settings tab.

Installed cleanly when we ran it

npm packageJavaScript1mo ago

AI review

dsh-pocket

What it is — Puts a phone-access tab in DSH settings that serves a LAN or cloudflared QR code so the phone can mirror and operate the desktop Web UI live.

Who it is for — The office machine is still running an agent while you are already on the commute, and you need the phone for the stream, a follow-up prompt, or an approval tap. Anyone who only drives DSH from the terminal and never starts dsh web has no surface for this plugin to attach to.

Watch out — It installed into a fresh profile in the sandbox and the harness registered it. dsh web has to be restarted before the phone-access tab shows up; turning on public access downloads cloudflared and the plugin runs shell commands locally. After a dsh web restart the phone must type the password again.

The verdict — LAN access stays enabled on my machine; the public tunnel stays down until I have actually left—keeping the full Web UI on a tunnel all day is not a trade I want.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Details
1.5k

Let your DSH agent browse Bilibili on its own, read subtitles, write a daily digest, and @ you when it finds something worth your attention.

Installed cleanly when we ran it

npm packageJavaScript8d ago

openpets

OpenPetsHQ

1.3k

Bridges DeepSeek Harness lifecycle status, errors, and approval requests to a locally running OpenPets desktop companion.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourceTypeScript1mo ago

AI review

openpets

What it is — Pushes DeepSeek Harness run state, errors, and pending approvals onto an OpenPets pet already open on the local desktop.

Who it is for — You are shipping code with a DeepSeek Harness agent on your own machine, and run state, failures, and approval prompts only flicker inside the session, so you want a desktop pet to catch them. Skip it when DSH lives on a headless remote host; nothing on that box can receive the events.

Watch out — There is no one-command package; you have to attach it as a subdirectory plugin from the repository. It invokes the system shell at runtime.

The verdict — I would not treat it as an agent feature; it does not change how I approve work, only whether I catch the prompt in peripheral vision.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Source

Connects DeepSeek Harness to Avernet's Bot Collaboration Network over WebSocket V2, with automatic onboarding, isolated agent sessions, tool-call events, and multi-bot routing tools.

No one-line install — this plugin lives inside a larger repository and publishes no npm package.

GitHub sourcePython1mo ago

AI review

deepseek-harness-channel-bcn

What it is — This plugin connects DeepSeek Harness to Avernet's Bot Collaboration Network.

Who it is for — When you need to run multi-agent collaboration tasks with DeepSeek Harness. If your work involves only independent agents without needing the Bot Collaboration Network, you can use other approaches.

Watch out — Sandbox testing has not been performed. Static check indicates it will execute shell commands. No obvious issues found.

The verdict — I would install it if I need Avernet infrastructure integration, because it supports WebSocket V2 connection.

Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.Read the source
1Source

dsh-qqbot

zaofan-make

225

QQ official bot channel (enhanced fork of tencent-connect/dsh-qqbot): full agent chat in groups and DMs, rich media, an auto-tagged sticker library with AI image replies, join approval, mute and broadcast, scheduled wakeups, remote approvals over QQ, interactive cards, multi-account instances, extensions, member profiles, and token-saving value-scored wakeups.

Installed cleanly when we ran it

npm packageTypeScript15d ago