lark-tasklist-export

A Claude Code skill that exports Feishu / Lark tasklists into structured Markdown documents, with an optional interactive Dashboard.

What It Does

Given a Feishu tasklist name or link, this skill:

1. Locates the tasklist via Feishu Task API
2. Fetches all uncompleted tasks (basic fields)
3. Enriches each task with custom field values (priority, project, group, etc.)
4. Reverse-engineers custom field option mappings (Feishu has no API for this — the skill infers GUID-to-name mappings from task context and asks you to confirm)
5. Resolves assignee/follower names via Contacts API
6. Fetches subtask details for tasks that have them
7. Generates a structured Markdown document
8. (Optional) Generates a self-contained HTML dashboard with board / grouped / table views, filters, and search

Prerequisites

• Claude Code installed
• lark-cli configured with user authentication
• Required scopes: task:task:read, task:tasklist:read, contact:user.base:readonly

Installation

Copy this folder into your Claude Code skills directory:

cp -r lark-tasklist-export ~/.claude/skills/

Usage

In Claude Code, just say:

"帮我导出飞书任务清单 XXX" "把我的飞书任务清单整理成文档" "导出任务列表为 markdown"

Or provide a tasklist link directly.

After generating the Markdown, you'll be asked whether you also want an interactive Dashboard HTML.

Output Example

Markdown — A structured document with:

• Stats overview (due today/tomorrow, high priority count, etc.)
• Task table sorted by due date, with custom field columns
• Subtask details with completion status
• Appendix with field definitions and team member list

Dashboard — A single HTML file you can open in any browser:

• Three views: Kanban board, grouped by project, full table
• Filter by priority / project / group
• Search tasks
• Subtask expansion

File Structure

lark-tasklist-export/
├── SKILL.md                          # Core workflow (8 steps)
├── README.md                         # This file
└── references/
    ├── output-template.md            # Markdown output template
    ├── dashboard-template.md         # Dashboard generation guide
    └── dashboard-template.html       # Dashboard style/layout template

How It Works

Feishu's Task API returns custom field values as opaque GUIDs rather than human-readable names, and there's no API to retrieve the option definitions. This skill works around that limitation by:

1. Collecting all unique GUID values across all tasks
2. Inferring names from task context (titles, tags like 【Legal】 or 【HR】)
3. Presenting the inferred mapping as a table for you to confirm or correct
4. Proceeding only after you've verified the mappings

This approach works for any tasklist — no hardcoded mappings.

License

MIT