lark-drive

lark-drive is a skill for Claude Code, Codex from ddpie/lark-mcp-on-agentcore. It costs 192 tokens per session (7,073 once invoked), scanned A, original, MIT.

A Feishu cloud-drive tool for managing online files and folders. Feishu cloud drive is a shared online storage area, similar to Google Drive or Dropbox.

In plain words
What is it for?
Use it to organize cloud-drive folders, upload or download files, inspect links and tokens, manage file versions and access, or import local Word, Markdown, Excel, CSV, PowerPoint, and Base files.
Why use it?
It removes the need to handle uploads, copies, moves, permissions, versions, and resource links manually across separate operations.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to organize cloud-drive folders, upload or download files, inspect links and tokens, manage file versions and access, or import local Word, Markdown, Excel, CSV, PowerPoint, and Base files.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/ddpie/lark-mcp-on-agentcore/lark-drive
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.

Any agent
npx skills add ddpie/lark-mcp-on-agentcore --skill lark-drive
Clone the repo
git clone --depth 1 https://github.com/ddpie/lark-mcp-on-agentcore

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for lark-drive

README.md
[![agentmods](https://agentmods.dev/badge/skills/ddpie/lark-mcp-on-agentcore/lark-drive/github.svg)](https://agentmods.dev/skills/ddpie/lark-mcp-on-agentcore/lark-drive)
Your own site
<a href="https://agentmods.dev/skills/ddpie/lark-mcp-on-agentcore/lark-drive"><img src="https://agentmods.dev/badge/skills/ddpie/lark-mcp-on-agentcore/lark-drive/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for lark-drive

Your own site · 80×15
<a href="https://agentmods.dev/skills/ddpie/lark-mcp-on-agentcore/lark-drive"><img src="https://agentmods.dev/badge/skills/ddpie/lark-mcp-on-agentcore/lark-drive.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 192 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,073 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00192 $0.07073
Opus 5 $0.00096 $0.03537
Sonnet 5 $0.00038 $0.01415
Haiku 4.5 $0.00019 $0.00707

Measured 10d ago against content hash 89f9d55646ef, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

lark-drive 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 10d 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.

docker/skills/lark-drive/SKILL.md · 222 lines

How it starts

The opening of the file, as written. The whole thing — 222 lines — stays where its author put it; the contents beside it link to each section on GitHub.

drive (v1)

术语说明: 飞书云空间也常被称为"云盘"、"云存储"、"网盘"或"我的空间",这些说法通常指的是同一个产品,是飞书官方的云端文件存储与管理中心。

导入分流规则: 如果用户要把本地 Excel / CSV / .base 快照导入成 Base / 多维表格 / bitable,必须优先使用 lark_drive_import(type="bitable")。不要先切到 lark-baselark-base 只负责导入完成后的表内操作。

副本分流规则: 如果用户要复制在线文档、创建文档副本、把文档复制到另一个文件夹,必须使用 lark_invoke(tool_name="lark_drive_files_copy", ...)。不要用 lark_drive_export 下载后再 lark_drive_import 上传,也不要用 lark_docs_fetch + lark_docs_create 重建正文;导出/导入只用于本地文件转换或离线产物。

快速决策

  • 用户要复制文档 / 创建副本 / 另存为副本时,使用 lark_invoke(tool_name="lark_drive_files_copy", ...)。先用 lark_discover(query="drive.files.copy") 确认参数;如果来源是 wiki URL/token,先用 lark_drive_inspect() 获取底层 tokentype,不要把 wiki token 直接当 file_tokenparams.file_token 传源文档 token,data.folder_token 传目标文件夹 token,data.name 传副本名称,data.type 传源文件类型(如 docx / sheet / bitable / slides)。示例:lark_invoke(tool_name="lark_drive_files_copy", args={params: {"file_token":"<DOC_TOKEN>"}, data: {"folder_token":"<FOLDER_TOKEN>","name":"<COPY_NAME>","type":"docx"}})。如返回需要确认,向用户确认后在调用里加 _confirm=true 重试。
  • 用户要识别飞书 / doubao 云空间 URL 的类型和 token时,可以先按 URL 路径形态做轻量判断;当路径已明确指向 docx / sheet / bitable / slides / file / folder 等资源时,可直接提取对应 token/type。传入 wiki URL、需要识别标题或 canonical URL、URL/token 有歧义,或后续操作依赖底层真实资源时,再使用 lark_drive_inspect(url="<url>") 进行识别;具体用法、失败处理和边界见 lark_get_skill(domain="drive", section="inspect")
  • 高风险写操作(删除、公开权限修改、owner 转移、版本删除/回滚、批量移动/覆盖/同步)必须同时满足三个条件才执行:目标已解析为该操作可直接使用的执行对象,执行细节已明确到可直接调用命令(例如删除的 file-token/type、公开权限修改的共享范围、owner 转移的目标 owner、版本删除/回滚的 version id、移动/覆盖/同步的目标位置和冲突策略),且用户在本轮明确确认执行这些具体目标和执行细节。用户只说"删除没用的文件""开放/共享给大家""改成开放""覆盖/移动这些"只表示目标状态;先只读发现并列出候选、权限档位或执行方案,停止等待用户确认。
  • 用户要检查 / 治理文档权限、公开范围、链接分享、外部访问、复制下载权限、密级标签、owner 转移,或要"权限风险报告、收紧权限、申请查看 / 编辑权限、转移 / 批量转移 owner",必须先调用 lark_get_skill(domain="drive", section="workflow"),再按其中 Workflow Registry 进入 permission_governance workflow(lark_get_skill(domain="drive", section="workflow-permission-governance"))。
  • 用户要整理云盘 / 文件夹 / 文档库 / 知识库 / 个人文档库,或要"盘点目录结构、找出未归档/临时/重复/空目录、生成整理方案",必须先调用 lark_get_skill(domain="drive", section="workflow"),再按其中 Workflow Registry 进入 knowledge_organize workflow(lark_get_skill(domain="drive", section="workflow-knowledge-organize"))。默认只生成方案;创建目录、移动资源、申请权限都必须单独确认。
  • 用户要搜文档 / Wiki / 电子表格 / 多维表格 / 云空间(云盘/云存储)对象,优先使用 lark_drive_search()。自然语言里"最近我编辑过的"、"我创建的"(→ created_by_me=true,原始创建者语义)、"我负责/owner 的"(→ mine=true,owner 语义)、"最近一周我打开过的 xxx"、"某人 owner 的 docx" 等直接映射到扁平参数,避免手写嵌套 JSON。
  • 用户要获取文档评论列表时,优先使用 lark_drive_list_comments(url="<url>"),不要优先手写 lark_invoke(tool_name="lark_drive_file_comments_list", ...);具体使用方式先调用 lark_get_skill(domain="drive", section="list-comments")
  • 用户要根据文档评论定位正文位置,例如 根据评论 review 文档、根据评论内容回看文档、区分多处相同引用文本时,对于 docx 类型(file_type=docx)的文档支持通过 lark_drive_list_comments(need_relation=true) 返回评论位置,其他类型会静默忽略该参数;具体用法需要先调用 lark_get_skill(domain="drive", section="comment-location") 了解。
  • 用户给出 doubao.com 的云空间资源 URL/token,或明确提到豆包里的 file/folder/docx/sheet/bitable/wiki 资源时,仍按资源类型、URL 路径和 token 路由到本 skill;不要因为域名不是飞书而回退到 WebFetch。
  • 用户要把本地 .xlsx / .csv / .base 导入成 Base / 多维表格 / bitable,第一步必须使用 lark_drive_import(type="bitable")
  • 用户要把本地 .md / .docx / .doc / .txt / .html 导入成在线文档,使用 lark_drive_import(type="docx")
  • 用户要把本地 .pptx 导入成飞书幻灯片,使用 lark_drive_import(type="slides");当前 PPTX 导入上限是 500MB。
  • 批量执行 lark_drive_import 且目标是同一个位置(同一 folder_token、默认根目录,或同一 target_token)时,必须串行执行;不要并发导入到同一位置,服务端可能返回并发冲突错误。
  • 用户要在 Drive 里上传、创建、读取、局部 patch 或覆盖更新原生 .md 文件(不是导入成 docx),切到 lark_get_skill(domain="markdown")
  • 用户要比较原生 .md 文件的历史版本差异,或比较远端 Markdown 与本地草稿,切到 lark_get_skill(domain="markdown")lark_markdown_diff;需要版本号时先用 lark_drive_version_history()
  • 用户要查看、下载、回滚或删除文件的历史版本,使用 lark_drive_version_history()lark_drive_version_get()lark_drive_version_revert()lark_drive_version_delete();这组工具同时支持 user identity 和 bot identity。
  • 用户要把本地 .xlsx / .xls / .csv 导入成电子表格,使用 lark_drive_import(type="sheet")
  • 用户要在云空间(云盘/云存储)里新建文件夹,优先使用 lark_drive_create_folder()
  • 用户要查看某个文件有哪些可下载预览格式,或想下载 PDF / HTML / 文本 / 图片等预览产物,使用 lark_drive_preview()
  • 用户要获取某个文件的封面图,优先使用 lark_drive_cover();先 list_only=true 看规格,再选 spec 下载。
  • 用户要导出云文档时,优先使用 lark_drive_export(url="<文档 URL>", file_extension="<格式>");详细参数、Wiki token 和错误码处理见 lark_get_skill(domain="drive", section="export")
  • 用户要把本地文件上传到知识库 / 文档库里的某个 wiki 节点下时,仍然使用 lark_drive_upload(wiki_token="<wiki_token>");不要误切到 wiki 域命令。
  • lark-base 只负责导入完成后的 Base 内部操作(表、字段、记录、视图),不要在"本地文件 -> Base"这一步提前切到 lark-base
  • 用户给的是 wiki URL / token,且后续还没明确底层资源类型时,先用 lark_drive_inspect() 解包;lark_drive_inspect() 失败后不要自动切到别的写接口继续尝试,先按错误提示处理权限、scope 或链接问题。
  • lark_drive_inspect() / lark_drive_upload() 遇到 not foundpermission deniedmissing scope 时,默认停止重试;只有 rate limit 或临时网络错误才适合有限重试。

Read the full file on GitHub · 222 lines

Files

What ships with it

41 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 10d ago First seen · 222 lines · 192 tokens per session scan A 89f9d55646ef

Subscribe to this mod's changes

lark-drive is a skill published in the GitHub repository ddpie/lark-mcp-on-agentcore (8 stars, last pushed 12d ago), licensed MIT. It adds 192 tokens to every session and 7,073 once invoked, about $0.0010 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.