dsh-feishu-gateway：把 DeepSeek Harness agent 接到飞书
`@kriskwok/dsh-feishu-gateway` 是一个将飞书消息路由至 DeepSeek Harness (DSH) agent 的插件，旨在让用户无需公网 Webhook 即可通过飞书私聊或群聊与本地 DSH 进行交互。该插件利用飞书长连接接入，支持持久会话管理，飞书对话或群 Topic 可映射为独立 DSH Session，并支持通过 `/new` 或自然语言指令开启新会话。功能上，它提供 Markdown 富文本回复、原生 Typing 状态提示、长任务流式进度卡片，以及基于点击按钮的审批和问答交互。Web 端特有的交互组件在飞书端会自动降级为可读提示。使用前需配置 DSH
Read Moredsh-feishucard: Bridging Feishu Streaming Cards for DeepSeek Harness
`dsh-feishucard` is a Feishu bridge plugin designed for DeepSeek Harness, aimed at solving the problem of robot message sending/receiving and conversation management in environments without a public IP. This plugin is based on the WebSocket long connection of Feishu's official SDK, allowing it to be used locally or on an intranet without requiring a domain name or tunnel. Core features include: real-time streaming display of Agent replies, tool call collapsible panels, and process utterances via Feishu card streaming; maintaining an independent and persistent Agent session for each chat, with support for commands such as `/new` and `/switch`; providing a `feishu_send` proactive message-sending tool and approval interaction cards (including an automatic timeout rejection mechanism); supporting multi-robot configuration, emoji receipts, and process keep-alive auto-restart. Installation requires configuring Feishu Open Platform permissions and handling dependencies.
Read MoreDeepSeek Harness Lark Bridge: Controlling DSH Sessions from Feishu/Lark
dsh-lark-bridge is a bidirectional control plugin for Lark/Feishu within the DeepSeek Harness (DSH) ecosystem, designed to link Lark conversations with DSH Sessions for remote viewing and control. This tool supports initiating, continuing, correcting, stopping, or resuming DSH tasks through Feishu, and maps Lark groups to specific DSH projects, with each topic thread corresponding to an isolated Session. Core functionalities include real-time task status updates via native cards, approval of tool calls, responses to agent questions, and secure interaction handling for text, images, and workspace files. It employs WebSocket for persistent connections, eliminating the need for public webhooks, and requires Node.js 22+ and pnpm for its runtime environment. Installation is done using `pnpm
Read More