lark-base

Manage Feishu Base tables, fields, records, views, dashboards, and workflows via lark-cli.

3|2|Updated Jul 21, 2026
One-click install
npx skills add https://github.com/zlsbksdxl/codex-lark --skill lark-base-zlsbksdxl
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lark-base
Source: https://github.com/zlsbksdxl/codex-lark/tree/main/plugins/feishu2codex/skills/lark-base
Command: npx skills add https://github.com/zlsbksdxl/codex-lark --skill lark-base-zlsbksdxl

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires lark-cli, and includes references (resource) components.

What problem does it solve? Working with Feishu/Lark Base (Bitable) through raw APIs requires resolving tokens, constructing complex JSON payloads, and handling pagination, permissions, and error codes manually. This Skill routes every Base operation through structured lark-cli shortcuts so agents can create tables, write records, run aggregations, and configure dashboards without guessing API shapes. ## Core Features & Use Cases - Table, Field & Record Management: Create and copy Bases, manage tables and fields (including formula and lookup fields), and upsert, batch-create, or batch-update records with correct CellValue formats. - Query & Aggregation: Run filtered, sorted record queries and cloud-side aggregations with +data-query, following a strict SOP that prevents drawing global conclusions from partial pages. - Dashboards, Forms, Workflows & Roles: Build dashboard chart blocks with validated data_config, submit and manage forms, configure workflows, and manage advanced permissions and custom roles. - Use Case: A user shares a /base/ link and asks for a monthly sales trend chart. The Skill resolves the base token via +url-resolve, reads the real table and field structure, creates a dashboard, and adds a line chart block with the correct data_config. ## Quick Start Ask the agent to list the tables and fields in a Feishu Base by sharing its /base/ link, then request a summary count of records grouped by status.

Frequently Asked Questions about lark-base

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

FAQPage Schema
How do I create or update records in a Feishu Base table?

Use lark-cli base +record-upsert, +record-batch-create, or +record-batch-update after reading the field structure with +field-list. Cell values must match each field type, such as arrays of {id} objects for user and link fields, and batches are limited to 200 records per call.

How do I run aggregation or Top N queries on Bitable data?

Use lark-cli base +data-query with a DSL containing dimensions, measures, filters, sort, and pagination so aggregation runs in the Base cloud service. Local jq or shell processing is only acceptable for results under 200 rows, not for global statistics.

How do I get the base_token from a Feishu Base URL?

Run lark-cli base +url-resolve --url "<url>" --as user to extract the real base_token and related IDs. Never pass a full URL, wiki token, or workspace token directly as --base-token, since that causes invalid token errors.

Can I import an Excel or CSV file into Feishu Base with this skill?

No, file import is out of scope for this skill. Use lark-cli drive +import --type bitable to convert Excel, CSV, or .base files into a Base first, then return to base commands for table and record operations.

Why does creating a formula or lookup field fail in Lark Base?

Formula and lookup fields require reading the formula-field-guide or lookup-field-guide references first, then passing --i-have-read-guide with +field-create or +field-update. The CLI fails fast without this flag, and expressions must use exact table and field names from +table-get.

What should I do when Base commands return permission errors like 91403?

Error 91403 means the identity lacks access to that Base; do not retry in a loop. Route to the lark-shared skill for authorization and scope recovery, and only retry with --as bot once when user identity fails without a recovery hint.