Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

yuqueyuque 文档

Agent Skill

yuque 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

605

周安装

26

GitHub Stars

16

下载量

212
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:yuque(yuque 文档)
来源仓库:https://github.com/flc1125/skills
仓库路径:skills/yuque
安装命令:
npx skills add https://github.com/flc1125/skills --skill yuque
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/flc1125/skills --skill yuque

简介

yuque 用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词快速定位候选结果。
  • 通过 npx skills add 命令从指定仓库安装,需结合原始 README 核验具体用法。
  • 安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。
  • yuque 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Yuque

Operate on Yuque through its OpenAPI with full capability awareness and explicit safety gates.

Operating Mode

Act as a Yuque workspace operator, not as a generic REST client.

Prioritize:

  • readable resource identification before raw IDs
  • read and verify before write
  • smallest effective change before broad updates
  • guarded handling for destructive operations
  • repo and TOC context before moving documents

Assume the upstream API base URL is https://www.yuque.com and authentication uses X-Auth-Token unless local evidence shows otherwise.

Read docs/auth.md when wiring local authentication for execution.

Capability Model

Treat Yuque operations as three execution classes.

1. Read

Default allow.

Includes:

  • current user and heartbeat
  • user groups
  • team members
  • repo lists and repo details
  • doc lists and doc details
  • TOC reads
  • search
  • doc versions
  • statistics

2. Safe Write

Default allow when the requested target is clear.

Includes:

  • create repo
  • update repo metadata
  • create doc
  • update doc
  • change member role
  • create TOC node
  • rename TOC node
  • move a doc or node within TOC without deleting anything

3. Dangerous Write

Default deny for execution. Produce a preflight plan unless the surrounding environment defines an explicit approval path.

Includes:

  • delete doc
  • delete repo
  • remove TOC node
  • remove group member
  • broad or implicit TOC rewrites
  • batch operations that can orphan or hide content

For dangerous requests:

  • identify the exact targets
  • explain impact and reversibility
  • list the endpoint and action that would be used
  • stop at the plan unless the user explicitly wants the higher-risk path and the current environment supports that confirmation model

Resource Identification

Prefer readable paths over opaque identifiers.

Prefer this order when the user does not force a specific form:

  1. group_login + book_slug + doc_slug
  2. group_login + book_slug + doc_id
  3. book_id + doc_id
  4. raw numeric IDs only when that is all that is available

Use the slug-based repo routes when the user speaks in human-readable Yuque paths. Use the ID routes when an existing integration already has stable IDs.

Before any TOC write, read the current TOC first to recover node UUIDs and parent-child relationships.

Core Workflows

Follow the smallest workflow that satisfies the request.

Read Workspace State

Use for questions such as:

  • "读取这个知识库"
  • "看看这个文档内容"
  • "列出这个团队的知识库"
  • "把当前目录结构给我理一下"

Sequence:

  1. Resolve the target repo or doc from readable identifiers.
  2. Read the resource using the narrowest matching endpoint.
  3. Return the fields that matter for the task.
  4. For docs, prefer body plus the relevant format-specific body field when present.

Read references/user-group.md for user, group, and search tasks. Read references/repo.md for repo tasks. Read references/doc.md for document tasks. Read references/toc.md for TOC tasks.

Create a Document

Use for requests such as:

  • "在这个知识库里新建文档"
  • "把这份 Markdown 发到语雀"
  • "在某个目录下面创建一篇文档"

Sequence:

  1. Resolve the target repo.
  2. If placement matters, read TOC first and resolve the target parent node.
  3. Build the create payload with explicit title, body, and optional slug, public, format.
  4. Default to format=markdown unless the user clearly requires html or lake.
  5. After create succeeds, place the document in TOC if the user requested a specific location.

Update a Document

Use for requests such as:

  • "修改这篇语雀文档"
  • "把这段内容追加到现有文档"
  • "把这个标题和 slug 改掉"

Sequence:

  1. Read the current doc first.
  2. Determine whether the user wants a full replacement or a targeted edit.
  3. Produce the updated full body content.
  4. Preserve fields the user did not ask to change, especially slug, public, and format.
  5. Execute the update against the matching repo-scoped route.

Do not treat doc editing as a blind patch. The API is safer to use when the resulting full document state is explicit.

Reorganize TOC

Use for requests such as:

  • "把这篇文档放到某个菜单下面"
  • "把这个节点移动到另一个父节点下"
  • "新建一个目录节点"
  • "重命名目录节点"

Sequence:

  1. Read the current TOC.
  2. Resolve the target node UUIDs and the intended parent or sibling relationship.
  3. Choose the narrowest non-destructive action:

- appendNode - prependNode only when head insertion is clearly intended - editNode

  1. Execute only the local TOC change needed for the request.
  2. Re-read TOC or summarize the expected new placement.

Do not use removeNode in normal execution.

Decision Rules

  • Prefer markdown for authoring unless the user clearly asks for html, lake, table, or sheet semantics.
  • Prefer repo-scoped doc endpoints over generic doc-detail routes when repo context is known.
  • Read before write for docs, repos, member changes, and all TOC actions.
  • When moving content in TOC, operate on node UUIDs recovered from a fresh TOC read.
  • Preserve visibility and slug unless the user explicitly asks to change them.
  • Treat missing target clarity as a blocker for writes; resolve the target first instead of guessing.
  • When the request spans many resources, break it into discrete operations and state the plan.

API Surface Map

This skill covers the major Yuque API domains:

  • user
  • search
  • group members
  • repo read and write
  • doc read and write
  • doc versions
  • TOC read and safe reorganization
  • statistics

Read references/capability-map.md for endpoint coverage and safety classification.

Prepare

Before executing authenticated scripts, prepare local auth like this:

  1. Copy skills/yuque/.env.example to skills/yuque/.env
  2. Replace the placeholder YUQUE_TOKEN with a real token
  3. Keep skills/yuque/.env local only and do not commit it

The scripts auto-load skills/yuque/.env when it exists.

Read docs/auth.md for token resolution order and auth details.

Scripts

Use the bundled scripts when a concrete request should turn into a reproducible Yuque API call.

  • scripts/heartbeat.mjs and scripts/current-user.mjs: inspect basic auth reachability and token identity
  • scripts/search.mjs: search docs or repos
  • scripts/list-user-groups.mjs, scripts/list-group-members.mjs, scripts/update-group-member-role.mjs: inspect groups and manage non-destructive membership changes
  • scripts/list-repos.mjs, scripts/read-repo.mjs, scripts/create-repo.mjs, scripts/update-repo.mjs: operate on repos
  • scripts/list-docs.mjs, scripts/read-doc.mjs, scripts/create-doc.mjs, scripts/update-doc.mjs: operate on docs
  • scripts/list-doc-versions.mjs, scripts/read-doc-version.mjs: inspect doc history
  • scripts/read-toc.mjs, scripts/update-toc.mjs, scripts/move-toc-node.mjs: inspect and reorganize TOC without deletion
  • scripts/read-statistics.mjs: inspect team statistics
  • scripts/*-preflight.mjs: prepare guarded plans for dangerous actions without executing them

Script rules:

  • default to preview mode
  • pass --execute only when the target is clear and the action is allowed by this skill's safety model
  • prefer a local skills/yuque/.env copied from .env.example for normal execution
  • allow --token to override local config when a one-off token is necessary
  • read the current doc or TOC before using write scripts

Error Handling

Use these defaults:

  • 400: request shape is wrong; re-check path params, query params, and enums
  • 401: token missing, invalid, or lacks scope
  • 403: token is valid but cannot access the target resource
  • 404: repo, doc, group, or node target is wrong
  • 422: payload validation failed; inspect required fields and enum values
  • 429: rate limited; reduce request volume and retry cautiously
  • 500: upstream failure; avoid repeating broad writes until the target state is checked again

Read references/auth-and-errors.md for auth and failure handling.

Dangerous Requests

When the user asks for deletion or removal:

  1. Do not execute the destructive call by default.
  2. Produce a preflight summary with:

- target identifiers - likely affected repo, doc, or nodes - endpoint and method - reversibility concerns

  1. Suggest a safer alternative when one exists.

Examples:

  • archive by renaming or moving a doc instead of deleting it
  • move a TOC node out of the way instead of removing it
  • restrict access or change role instead of removing a member

Read references/safety-policy.md for the full policy.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.86%
按下载量换算80

Claude

29.22%
按下载量换算62

Cursor

19.39%
按下载量换算41

Gemini CLI

10.27%
按下载量换算22

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills