dsh-feishu-bot
Run the following command in DeepSeek Harness:
dsh plugin install zyfgood/dsh-feishu-bot
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install zyfgood/dsh-feishu-bot in the DeepSeek Harness terminal to install the plugin; source code is available at https://github.com/zyfgood/dsh-feishu-bot .
About this plugin
dsh-feishu-bot bridges DeepSeek Harness agents into Feishu / Lark, solving a practical problem: you no longer need a laptop open to drive an agent—send a message from your phone and get streamed answers back. The connection uses the official Larksuite SDK WebSocket channel, so there is no public callback URL to configure and auto-reconnect is built in.
The plugin covers full bidirectional interaction. On the inbound side, three reply modes are available (echo for binding verification, llm for direct model responses with per-chat memory, and agent for handing messages to a dedicated DSH agent session with the standard preset and full tooling). Replies stream live into Feishu using native typewriter cards. If a task is already running, a mid-flight message is injected via steer at the next step boundary—the agent answers you first, then resumes. On the outbound side, nine feishu_* tools let the agent send messages, post interactive cards, query sessions, and push results to a configured group channel. When a decision is needed, a button card appears in the chat: click a button, reply with a number, or type the option text; a timeout guard prevents the session from hanging. Session-management commands work right in the chat, and session-to-agent mappings persist across restarts so context is automatically resumed.
It is built for teams that live in Feishu: developers who want to direct an agent during a commute or a meeting, engineers who need task results pushed into a group channel for collaborative notification, and integrators who want a zero-infrastructure way to verify LLM plumbing end-to-end.
Use Cases
- Send a message from Feishu on your phone to drive code generation and task execution remotely
- Agent pushes completed task results to a Feishu group for team notification
- Quickly verify LLM binding end-to-end from your phone during a meeting break
Best For
- AI application development teams working within the Feishu ecosystem
- Engineers who want to drive DeepSeek Harness agents remotely
- Integration engineers needing a low-friction way to verify LLM pipelines
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.