feishu-mcp-manual

星标9

分支13

更新时间2026年7月19日 08:12

Progressive-disclosure usage manual for the Feishu (Lark) MCP tool. Read this when you need detail beyond the one-line action descriptions: receive_id vs receive_id_type (open_id/chat_id), send vs reply, check/read/search, placeholder + edit for long responses, contacts/accounts basics, the notification transient-hook vs persistent-context split, and side-effect caveats. Pulled on demand via action='manual'; you do not need to call it before every send.

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

来源

Lingtai-AI

Lingtai-AI/lingtai-kernel

打开 GitHub 仓库查看创作者相关仓库

下载

Feishu (Lark) MCP — usage manual (progressive disclosure)

RECIPIENTS: receive_id / receive_id_type

send targets a recipient by receive_id plus receive_id_type. Use receive_id_type='open_id' for an individual user (ou_xxx) and receive_id_type='chat_id' for a group chat (oc_xxx). receive_id_type defaults to open_id when omitted.
email, user_id, and union_id are also accepted as receive_id_type values when you only have that identifier for a user.

SEND vs REPLY

reply (message_id from read/check results, text) threads your response to a specific incoming message; prefer it when answering a particular message.
send (receive_id, receive_id_type, text) starts a fresh message; use it for unsolicited or standalone messages.

READING: check / read / search

check: list recent conversations with unread counts (optional account).
read: read messages from one chat (chat_id; optional limit, account).
search: regex search over inbox messages (query; optional account, chat_id).

PLACEHOLDER / PROGRESS

For responses that take more than ~5s, send action='send' with placeholder=true and your interim text. This returns a compound message_id.
Update it later with action='edit', message_id=<that id>, text=<final> instead of sending a second message, so the user sees one evolving reply.

CONTACTS / ACCOUNTS

contacts: list saved contacts (optional account).
add_contact: save a contact alias (open_id, alias; optional name, chat_id). Saving an alias does not grant inbound permission on its own.
remove_contact: remove a contact (alias or open_id).
accounts: list configured app accounts.

MESSAGE IDS

message_id is the compound id returned by read/check ({alias}:{chat_id}:{feishu_message_id}); pass it back verbatim to reply/edit/delete.

NOTIFICATIONS: TRANSIENT HOOK vs PERSISTENT CONTEXT

Inbound Feishu messages surface to the agent in two _meta lanes:

_meta.agent_meta.notifications.attention.mcp.feishu — a compact high-attention hook only: data.message_ids and dismiss guidance, never message text or routing context.
_meta.agent_meta.notifications.persistent.mcp.feishu — durable context: recent conversation messages (bounded text, both directions), sender/chat routing hooks, reply refs when present, and per-message comments for the agent's own outgoing messages or truncated text.

The feishu tool remains the source of truth: neither lane marks anything read, so use read/check for exact producer state — especially when a persistent message is truncated. Reply in Feishu when the message arrived through Feishu (reply with the compound message id, or send to the chat/open_id). After handling, dismiss the transient hook with notification.dismiss_channel("mcp.feishu"); the persistent block is context history, not unread state. Generic mirror-vs-canonical-state and dismiss-safety rules live in notification-manual.

SIDE EFFECTS & ERROR SURFACING

send, reply, and edit deliver to real users — they are external side effects, so confirm recipient and content before sending unsolicited messages.
Actions return a result dict on success or {'error': <message>} on failure (e.g. missing receive_id, bad message_id). Check for the 'error' key and surface or act on it rather than assuming delivery.

同仓库更多 Skills

同仓库

lingtai-doctor

Lingtai-AI/lingtai-kernel

Read-only health diagnostics for LingTai agents and bots. Use when an agent appears offline or unreachable, when a machine migration may have left stale MCP/addon command paths, when heartbeat/status/process/notification/log surfaces disagree, or before deciding whether to mail, refresh, CPR, or edit persistent configuration. Includes a bundled doctor.py script for layered local checks without exposing secrets.

2026-07-199

environment-variables

Lingtai-AI/lingtai-kernel

Complete catalogue of environment variables consumed by the LingTai kernel, wrapper, bundled MCPs, adapters, tests, and daemon composition surfaces.

2026-07-199

system-manual

Lingtai-AI/lingtai-kernel

Second-layer router for LingTai's progressive-disclosure operating manuals. Read this when resident substrate/procedures are too compact and you need the right lower reference. Route here for the expanded substrate/runtime model, lifecycle and `system` tool actions, the init.json composition and preset runtime model route, action/procedure discipline and skill routing, tool-result summarization, SQLite/`log.sqlite` trace inspection and trajectory mining, runtime/kernel update checks and nudges, environment variables, goal notifications, molt/memory questions, MCP/addon ownership, collaboration topology, and resident prompt design. The nested catalog below names each reference and its exact trigger.

2026-07-199

lingtai-kernel-dev

Lingtai-AI/lingtai-kernel

Mandatory repository-local development guide for LingTai's Python kernel and runtime. Use this before changing code, architecture documents, tests, packaging, capabilities, adapters, or developer documentation in lingtai-kernel. Routes each task through the exact baseline, the distributed ANATOMY/CONTRACT systems, focused validation, and pull-request safety gates.

2026-07-199

file-manual

Lingtai-AI/lingtai-kernel

Operational guide for LingTai's built-in file tools: read, write, edit, glob, and grep. Use when working with local text files, deciding whether to use file tools versus bash, handling large files, avoiding binary/image misuse, or reading non-UTF-8 text via explicit bash/Python/iconv instead of complicating the core read tool. Covers UTF-8 policy, safe write/edit discipline, search workflows, and examples for GBK/Shift-JIS/Latin-1 conversion.

2026-07-199

lingtai-kernel-anatomy

Lingtai-AI/lingtai-kernel

Router for LingTai's distributed code navigation system. Use this before navigating, creating, or maintaining kernel ANATOMY.md files. The normative anatomy-of-anatomy now lives at the repository-root ANATOMY.md; this skill explains how an agent enters and descends that graph, how it pairs with the distributed CONTRACT.md interface-definition graph, and what to do when code and navigation disagree.

2026-07-199

name	feishu-mcp-manual
description	Progressive-disclosure usage manual for the Feishu (Lark) MCP tool. Read this when you need detail beyond the one-line action descriptions: receive_id vs receive_id_type (open_id/chat_id), send vs reply, check/read/search, placeholder + edit for long responses, contacts/accounts basics, the notification transient-hook vs persistent-context split, and side-effect caveats. Pulled on demand via action='manual'; you do not need to call it before every send.
version	1.2.0
last_changed_at	"2026-07-19T00:00:00.000Z"
related_files	["src/lingtai/mcp_servers/feishu/manager.py","src/lingtai/mcp_servers/feishu/server.py","src/lingtai/mcp_servers/feishu/service.py"]
maintenance	Tracks the MCP server's manager/config behavior; update when the server's setup or API surface changes.

Feishu (Lark) MCP — usage manual (progressive disclosure)

RECIPIENTS: receive_id / receive_id_type

send targets a recipient by receive_id plus receive_id_type. Use receive_id_type='open_id' for an individual user (ou_xxx) and receive_id_type='chat_id' for a group chat (oc_xxx). receive_id_type defaults to open_id when omitted.
email, user_id, and union_id are also accepted as receive_id_type values when you only have that identifier for a user.

SEND vs REPLY

reply (message_id from read/check results, text) threads your response to a specific incoming message; prefer it when answering a particular message.
send (receive_id, receive_id_type, text) starts a fresh message; use it for unsolicited or standalone messages.

READING: check / read / search

check: list recent conversations with unread counts (optional account).
read: read messages from one chat (chat_id; optional limit, account).
search: regex search over inbox messages (query; optional account, chat_id).

PLACEHOLDER / PROGRESS

For responses that take more than ~5s, send action='send' with placeholder=true and your interim text. This returns a compound message_id.
Update it later with action='edit', message_id=<that id>, text=<final> instead of sending a second message, so the user sees one evolving reply.

CONTACTS / ACCOUNTS

contacts: list saved contacts (optional account).
add_contact: save a contact alias (open_id, alias; optional name, chat_id). Saving an alias does not grant inbound permission on its own.
remove_contact: remove a contact (alias or open_id).
accounts: list configured app accounts.

MESSAGE IDS

message_id is the compound id returned by read/check ({alias}:{chat_id}:{feishu_message_id}); pass it back verbatim to reply/edit/delete.

NOTIFICATIONS: TRANSIENT HOOK vs PERSISTENT CONTEXT

Inbound Feishu messages surface to the agent in two _meta lanes:

_meta.agent_meta.notifications.attention.mcp.feishu — a compact high-attention hook only: data.message_ids and dismiss guidance, never message text or routing context.
_meta.agent_meta.notifications.persistent.mcp.feishu — durable context: recent conversation messages (bounded text, both directions), sender/chat routing hooks, reply refs when present, and per-message comments for the agent's own outgoing messages or truncated text.

SIDE EFFECTS & ERROR SURFACING

send, reply, and edit deliver to real users — they are external side effects, so confirm recipient and content before sending unsolicited messages.
Actions return a result dict on success or {'error': <message>} on failure (e.g. missing receive_id, bad message_id). Check for the 'error' key and surface or act on it rather than assuming delivery.