lark-base

A tool for managing Feishu Base, a cloud database made from tables, fields, records, and views. It also covers calculations, forms, dashboards, workflows, and access roles.

In plain words
What is it for?
Use it to create and edit tables, fields, records, views, formulas, reports, forms, dashboards, workflows, and permissions in Feishu Base.
Why use it?
It lets an agent manage structured business data inside Feishu without confusing it with ordinary documents or local files. It supports both data changes and the surrounding Base setup.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/dropfan/claude-code-plugins/lark-base
Any agent
npx skills add DropFan/claude-code-plugins --skill lark-base
Clone the repo
git clone --depth 1 https://github.com/DropFan/claude-code-plugins

Made for: Claude Code, Codex.

Per session 84 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,897 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00084 $0.06897
Opus 5 $0.00042 $0.03449
Sonnet 5 $0.00017 $0.01379
Haiku 4.5 $0.00008 $0.00690

Measured 2d ago against content hash b6e9cc0c9c02, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

lark-base scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/lark/skills/lark-base/SKILL.md · 187 lines

How it starts

The opening of the file, as written. The whole thing — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.

⚙️ Cowork / Claude Desktop 执行环境说明(自动注入)

本技能依赖本地 lark-cli@larksuite/cli,可用 command -v lark-cli 定位)及其 ~/.lark-cli 登录态(应用密钥存于 macOS keychain)。

在 Cowork 中运行任何 lark-cli 命令时,必须在本地 macOS 上执行(使用 Desktop Commander 的 start_process / interact_with_process,或其它本地 shell 工具),不要用隔离的 Linux 沙箱 mcp__workspace__bash——沙箱里没有 lark-cli、也读不到 keychain。 执行前确保 npm 全局 bin 目录(npm prefix -g 输出目录下的 bin)在 PATH 中。

(在 Claude Code 中可忽略本说明,lark-cli 在本机 shell 直接可用。)

base

何时使用

使用本 skill:

  • 用户明确提到 Base / 多维表格 / bitable,或给出 /base/ 链接。
  • 用户要在 Base 内建表、改表、管理字段、写记录、查记录、配视图。
  • 用户要在 Base 内做公式字段、lookup 字段、跨表计算、派生指标、筛选聚合、TopN、统计分析。
  • 用户要管理 Base 表单、仪表盘、workflow、高级权限或角色。
  • 用户要把旧 Base 聚合式命令或旧写法迁移到当前 lark-cli base +... shortcut。

不要使用本 skill:

  • 只是认证、初始化配置、切换身份、处理 scope 或权限授权恢复,转 lark-shared
  • 把本地文件导入成 Base,或将 Base 导出为本地文件,转 lark-drive
  • 泛化数据分析、字段设计、公式讨论,但没有 Base/多维表格上下文。

使用边界

  • Base 业务操作只使用 lark-cli base +... shortcut,不使用旧聚合式 +table / +field / +record / +view / +history / +workspace
  • 执行 update 前必须先查当前 shortcut 的 --help 或对应 reference。若命令要求完整配置,首次请求必须基于可信的当前配置执行 read-modify-write:只修改用户明确指定的内容,保留其他仍适用的可写配置,并按命令要求的结构提交。若命令支持局部/delta update,按其契约提交最小合法 payload;不得以不完整请求试错补参。
  • Base CLI/OpenAPI 当前不支持视图行高、冻结列、列宽等 UI-only 外观设置。遇到这类需求,说明能力边界并停止,不要猜测未文档化参数或改走 raw API。
  • 本地文件与 Base 之间的导入/导出转 lark-drive,具体格式、参数、路径限制和仅结构导出规则由 lark-drive 负责;导入完成后再回到 Base 命令。
  • 在线复制 Base 使用 +base-copy,不要绕行导出/导入。
  • 认证、初始化、scope、身份切换、权限不足恢复属于 lark-shared;Base 文档只保留会影响 Base 路径选择的权限规则。

先获取 Base Token 和所需 ID

进入任何需要目标 Base 的 shortcut 前,必须先拿到可用的 base_token,以及当前任务需要的 table_id / view_id / record_id / form_id / dashboard_id / workflow_id 等真实 ID;不要把完整 URL、wiki token、workspace token 或孤立 raw token 直接当作 --base-token

  • 用户输入 URL 或分享链接:先运行 lark-cli base +url-resolve --url "<url>" --as user,用返回的 base_token 和相关 ID 继续后续命令。
  • Base/Wiki URL 的 table= query 参数实际表示当前选中的顶层 block,可能是数据表、仪表盘或 workflow;不要按参数名自行当成 table_id。以 +url-resolve 返回的 block_type 以及 table_id / dashboard_id / workflow_id 为准;selection_source=url_query 只说明 URL 当前选中了该 block,不代表它覆盖用户明确点名的目标。若用户点名的 dashboard 与 block_name 不一致,先用 +dashboard-list 按名称匹配;若只返回中性 block_id,按 hint 用 +base-block-list 确认类型。
  • 用户输入 Base 标题、关键词或不确定名称:先运行 lark-cli base +title-resolve --title "<keyword>" --as user--title 传入标题中的短关键词,不超过 30 个字符;过长标题先取最有区分度的短关键词;多候选时先让用户消歧,不要猜。
  • 文档嵌入 Base 标签:直接读取 <bitable> / <base_refer>token 作为 --base-tokentable-id 作为 --table-idview-id 作为 --view-id;孤立 raw token 不走 +url-resolve
  • 仍无法定位且用户不是要新建 Base 时,先反问用户要操作哪一个 Base;用户要新建时才用 +base-create

Read the full file on GitHub · 187 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 187 lines · 84 tokens per session scan A b6e9cc0c9c02

Subscribe to this mod's changes

lark-base is a skill published in the GitHub repository DropFan/claude-code-plugins (7 stars, last pushed 27d ago), licensed MIT. It adds 84 tokens to every session and 6,897 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.