AI Agent Hub
Back to plugins
🖥️

dsh-feishu-channel

Client Updated 2026.09.01

Run the following command in DeepSeek Harness:

dsh plugin install WyattJHayes/dsh-feishu-channel

Paste the following prompt into your AI chat to install this plugin:

Install the plugin in DeepSeek Harness by running dsh plugin install WyattJHayes/dsh-feishu-channel; the source is available at https://github.com/WyattJHayes/dsh-feishu-channel .

About this plugin

You want to drive a local DeepSeek Harness Agent from Feishu after leaving the terminal, without exposing any public port or running a reverse proxy. dsh-feishu-channel uses Feishu official WebSocket long-connection to receive events and replies via Open API, so nothing listens on a public port and the session survives laptop-lid closes or Wi-Fi switches.

After binding a project directory you reuse one Agent session shared on both the desktop client and the plugin side. Plain prompts run in FIFO order; stage summaries and the final result are posted back to Feishu automatically. Slash commands such as /new, /project, /status, /cancel, /approve, /deny, and /help cover day-to-day operations. High-risk tool calls file deletion, external writes, publish pushes, forced Git actions trigger a one-time approval with a ten-minute default expiry; cancellation or disconnection is treated as a deny, and the approval token is single-use within the originating chat only. All access is gated by an open_id and chat_id allowlist plus a project-root directory allowlist; group-chat output is addressed solely to allowlisted senders by default, and unauthorised events produce no reply to prevent identity probing.

Best suited to developers already running DeepSeek Harness who want to kick off read-only queries, code generation, or task approvals from a phone or the Feishu desktop app; and to small teams that need to share Agent output with a handful of trusted members while treating every allowlisted member as holding terminal-level privileges, tightening the blast radius with a restricted preset and a minimal set of allowedProjectRoots.

Use Cases

  • Kick off read-only queries or code tasks from the Feishu mobile app after leaving the terminal
  • Share Agent output with a small circle of trusted members via open_id allowlists in a team setting
  • Approve or deny high-risk operations such as file deletion and publish pushes directly in Feishu with single-use tokens

Best For

  • Independent developers already using DeepSeek Harness who want mobile access to their local Agent
  • Small teams that need to share Agent output with trusted members while tightly gating access
  • Security-conscious users who prefer no public ports and a restricted preset to limit privileges