lark-mail

Manage Lark email workflows including drafting, replying, and attachment handling.

1|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/eatbreads/.agents --skill lark-mail-eatbreads
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lark-mail
Source: https://github.com/eatbreads/.agents/tree/main/skills/lark-mail
Command: npx skills add https://github.com/eatbreads/.agents --skill lark-mail-eatbreads

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Provides a safe, consistent interface to read, triage, draft, edit, forward, reply to, search, and send emails in Lark (Feishu) while enforcing strong guardrails to prevent accidental or unauthorized sending and to mitigate prompt-injection attacks embedded in email content.

Core Features & Use Cases

  • Inbox Triage & Search: List and search messages by query, folder, label, date range, and sender to decide which items need action.
  • Read & Thread View: Read single messages or full threads with normalized plain-text and optional HTML bodies, plus attachment metadata.
  • Drafting & Sending Workflows: Create, edit (structured patch operations), forward, reply, reply-all, and send messages with enforced user confirmation before delivery.
  • Attachment & Inline Media Handling: Support for adding/removing attachments and inline CID images while preserving MIME structure.
  • Safety & Compliance: Explicit rules to treat email content as untrusted data, require schema inspection, require --confirm-send for sends, and mandate send_status checks after delivery.

Quick Start

Use the lark-cli mail +reply shortcut to create a draft reply for a specified message and do not send until the user explicitly confirms.

Frequently Asked Questions about lark-mail

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Lark email triage and inbox search securely?

Automate Lark email triage by searching messages through query, folder, label, date range, and sender filters to identify actionable items. The workflow treats all email content as untrusted data to mitigate prompt-injection attacks during inbox triage operations.

Can I draft and send Lark emails programmatically without accidental sends?

Draft and send Lark emails programmatically while enforcing explicit user confirmation before delivery. The workflow requires the --confirm-send flag and mandates send_status checks after delivery to prevent accidental or unauthorized email sends.

What do I need to manage Lark mail workflows via lark-cli?

Managing Lark mail workflows requires lark-cli availability and appropriate mail scopes including read, modify, send, and event permissions. These prerequisites enable programmatic mail operations like drafting, forwarding, and replying.

How does attachment and inline media handling work in Lark email drafts?

Attachment and inline media handling supports adding and removing files and inline CID images while preserving the MIME structure. This ensures attachments and inline media remain intact during draft editing and sending workflows.

Why treat email content as untrusted data when replying to Lark threads?

Treat email content as untrusted data when replying to Lark threads to mitigate prompt-injection attacks embedded in message bodies. Strict schema inspection and prompt-injection defenses prevent malicious content from triggering unauthorized actions.

Does lark mail support reply-all and forwarding with structured patch operations?

Lark mail supports reply-all and forwarding with structured patch operations for editing drafts. Users can create, edit, forward, and reply to messages while the system enforces confirmation rules before final delivery.