Skip to main content
Glama
Feishu (Lark) OpenAPI MCP Server

Feishu (Lark) OpenAPI MCP Server

by zhengjy01

dsh-feishu-mcp — Feishu (Lark) MCP connection for DeepSeek Harness

Connect DeepSeek Harness (DSH) to Feishu / Lark through the official @larksuiteoapi/lark-mcp server: the plugin spawns the official MCP server as a subprocess with your Feishu app credentials, discovers its tools, and registers them on the agent as mcp__feishu__* — so the DSH agent can send IM messages, read/write Bitable (multidimensional tables), manage docs, calendar, drive, and more, using your Feishu app's permissions.

A web settings panel (Settings → 飞书) configures the App ID / App Secret, shows connection status, and lists the discovered tools.

中文说明见 README.zh.md。

Features

  • Official MCP server: drives @larksuiteoapi/lark-mcp (Feishu OpenAPI MCP, Beta) over stdio — no reverse proxy, no public IP needed.

  • Full tool surface: every tool the official server exposes (per your app's permissions and -t presets) becomes an agent tool mcp__feishu__* — e.g. mcp__feishu__im_v1_message_create, mcp__feishu__bitable_v1_app_create.

  • App identity or user identity: configure appId/appSecret (tenant identity) and optionally a userAccessToken (user identity for private resources).

  • Agent tools: feishu_status / feishu_config / feishu_test / feishu_tools.

  • Web settings panel: configure credentials, test the connection, and see the live tool list.

Related MCP server: Lark MCP Custom

Compatibility

Requires DeepSeek Harness ≥ 0.1.5-rc.1 (declared as dsh.engines.dsh in the package manifest, so the DSH plugin marketplace can report it) and is verified against 0.1.5-rc.1. This build carries the DSH 0.1.5 adaptations: the strict tool-result contract (lossless-JSON snapshot, additionalProperties: false schema validation, and output.render returning ContentBlock[]) plus executable resolution that survives a launchd-started host whose PATH is only /usr/bin:/bin. Since this version the peerDependencies range explicitly declares compatibility with DSH 0.2.0-rc.2 (^0.2.0-rc.2 is now included), so no compatibility warning appears on 0.2.0-rc.2. No functional change.

Prerequisites

  1. A Feishu self-built app: Feishu Open Platform → Developer Console → create an app.

  2. Add the permissions you need (im:message, bitable:app, docx:document, calendar:calendar, drive:drive …) and publish a version.

  3. Install the plugin and configure the App ID / App Secret.

Install

# from npm (published package)
dsh plugin --profile web add dsh-feishu-mcp

# or local development
dsh plugin --profile web add link:/path/to/dsh-feishu-mcp
# restart dsh web to activate

Configure

Via the web settings panel (Settings → 飞书), or the agent tool:

feishu_config appId=<cli_xxx> appSecret=<your_secret>

Optional user identity (for private resources):

feishu_config userAccessToken=<user_access_token>

Usage

After connecting, the agent sees the Feishu tools as mcp__feishu__*. Verify:

feishu_test

Dev

pnpm install
pnpm build && node tests/smoke.mjs

License

MIT

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • -
      license
      Not graded
      quality
      Not graded
      maintenance
      A tool designed to help users connect AI Agents with the Feishu/Lark platform, encapsulating Feishu/Lark Open Platform API interfaces as MCP tools for document processing, conversation management, calendar scheduling and more.
      33,913 npm
      -
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to control Feishu (Lark) through structured tools for messaging, calendar, tasks, and documents.
      3
      MIT
    • A
      license
      Not graded
      quality
      F
      maintenance
      Enables AI agents to send and update Feishu/Lark messages, cards, files, and images, read recent chat context and block on human replies, and work with bitables, spreadsheets, and cloud docs through 25 tools. Runs over long connection or HTTP callbacks, so no public IP is needed.
      MIT