Token导航 LogoToken导航

Agent Skill · 研究检索

notion-connect(Notion connect 搜索)

用于处理 Notion 页面、数据库、工作区内容和结构化记录。它适合让 Agent 查询知识库、整理页面内容、创建记录或把外部信息同步到 Notion。使用时需要确认集成是否已被授权到目标页面或数据库,并区分读取、追加和覆盖更新;涉及批量写入或修改数据库属性时,应先核对字段名称、属性类型和目标页面。

notion-connect 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证为 MIT,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 52,401。

总安装 559周安装 24GitHub Stars 2下载 196核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

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:notion-connect(Notion connect 搜索)
来源仓库:https://github.com/abdullahbeam/nexus-design-abdullah
仓库路径:skills/notion-connect
安装命令:
npx skills add https://github.com/abdullahbeam/nexus-design-abdullah --skill notion-connect
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/abdullahbeam/nexus-design-abdullah --skill notion-connect

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于处理 Notion 页面、数据库、工作区内容和结构化记录。适合让 Agent 查询知识库、整理页面内容、创建记录或把外部信息同步到 Notion。使用时需要确认集成是否已被授权到目标页面或数据库,并区分读取、追加和覆盖更新。

安装命令:npx skills add https://github.com/abdullahbeam/nexus-design-abdullah --skill notion-connect。涉及批量写入或修改数据库属性时,应先核对字段名称、属性类型和目标页面。

来源仓库内路径:skills/notion-connect

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/abdullahbeam/nexus-design-abdullah --skill not…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

notion-connect 怎么安装?

在终端执行 npx skills add https://github.com/abdullahbeam/nexus-design-abdullah --skill notion-connec…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

notion-connect 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

notion-connect 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

notion-connect 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

notion-connect 用的人多吗?

公开数据是累计安装 559 次,近一周安装 24 次,下载 196 次,GitHub 2 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

notion-connect 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
notion-connect

Notion connect 搜索 · 本页

5592接触敏感数据MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Notion Connect

Meta-skill for complete Notion workspace integration.

Purpose

Enable natural language interaction with ANY Notion database. User says "query my Projects database" or "add a page to CRM" and it just works - no manual API calls, no remembering database IDs, no schema lookups.


Shared Resources

This skill uses notion-master shared library. Load references as needed:

ResourceWhen to Load
notion-master/scripts/check_notion_config.pyAlways first (pre-flight)
notion-master/references/setup-guide.mdIf config check fails
notion-master/references/error-handling.mdOn any API errors
notion-master/references/api-reference.mdFor API details

First-Time User Setup

If user has never used Notion integration before:

  1. Run config check to detect setup state: python 00-system/skills/notion-master/scripts/check_notion_config.py
  2. If exit code 2 (not configured), run the interactive setup wizard: python 00-system/skills/notion-master/scripts/setup_notion.py The wizard will:

- Guide user to get/enter API key - Save configuration to .env - Get user's Notion ID (for export functionality) - Save to user-config.yaml - Auto-run database discovery - Create 01-memory/integrations/notion-databases.yaml

  1. After setup completes, user can immediately start querying databases

Setup triggers: "setup notion", "connect notion", "configure notion"


Workflow 0: Config Check (ALWAYS FIRST)

Every workflow MUST start with config validation:

python 00-system/skills/notion-master/scripts/check_notion_config.py

Exit code meanings:

  • Exit 0: Fully configured (query, import, AND export all work)
  • Exit 1: Partial config (query and import work, export needs user ID)
  • Exit 2: Not configured (run setup wizard)

If exit code 2 (config incomplete):

  1. Tell user: "Notion integration needs to be set up first."
  2. Run: python 00-system/skills/notion-master/scripts/setup_notion.py
  3. Restart workflow after setup complete

If exit code 0 or 1: Continue to requested operation


Workflow 1: Discover Databases

Triggers: "connect notion", "sync notion", "discover databases", "what databases", "refresh notion"

Purpose: Find all accessible databases in user's Notion workspace and cache schemas.

Steps:

  1. Run config check (Workflow 0)
  2. Run discovery script: python 00-system/skills/notion-master/scripts/discover_databases.py
  3. Script outputs:

- Number of databases found - Database names and IDs - Creates/updates: 01-memory/integrations/notion-databases.yaml

  1. Show user summary of discovered databases
  2. Confirm context file saved

First-time flow: If notion-databases.yaml doesn't exist, discovery runs automatically.


Workflow 2: Query Database

Triggers: "query [database]", "find in [database]", "search [database]", "show [database]", "list [database]"

Purpose: Query any database by name with optional filters.

Steps:

  1. Run config check (Workflow 0)
  2. Load context: Read 01-memory/integrations/notion-databases.yaml

- If file doesn't exist → Run Workflow 1 (Discover) first

  1. Match database name (fuzzy):

- User says "Projects" → matches "Client Projects", "My Projects", etc. - If multiple matches → Show disambiguation prompt - If no match → Suggest running discovery

  1. Run query: python 00-system/skills/notion-master/scripts/search_skill_database.py --db <database_id> [--filter "..."] [--sort...] [--limit N]
  2. Format and display results using property types from cached schema
  3. Offer follow-up actions: "Want to add a page?" / "Query with different filters?"

Filter Syntax (load notion-master/references/filter-syntax.md if user needs help):

  • --filter "Status = Active"
  • --filter "Priority = High"
  • --filter "Tags contains Design"

Note: Currently supports single filter per query. Multiple filters (AND/OR) planned for future.


Workflow 3: Create Page

Triggers: "add to [database]", "create in [database]", "new [item] in [database]"

Purpose: Create a new page in any database with property validation.

Steps:

  1. Run config check (Workflow 0)
  2. Load context and match database (same as Workflow 2)
  3. Load schema for target database from context file
  4. Prompt user for required properties based on schema:

- Show property name + type + options (for select/multi-select) - Validate input against property type

  1. Run create: python 00-system/skills/notion-master/scripts/create_page.py --db <database_id> --properties '{"Title": "...", "Status": "..."}'
  2. Confirm creation with page URL
  3. Offer: "Add another?" / "View in Notion?"

Workflow 4: Manage Database Schema

Triggers: "create database", "new database", "add property to [database]", "update [database] schema"

Purpose: Create new databases or modify existing database schemas.

Steps:

Create New Database

  1. Run config check (Workflow 0)
  2. Get parent page ID (where database will live)
  3. Define database name and properties: python 00-system/skills/notion-master/scripts/manage_database.py create \ --parent <page_id> \ --title "Database Name" \ --properties '[{"name": "Name", "type": "title"}, {"name": "Status", "type": "select", "options": ["Todo", "Done"]}]'
  4. Confirm creation with database URL
  5. Run discovery to update context: python discover_databases.py

Update Schema (Add Property)

  1. Run config check
  2. Match database name (fuzzy) from context
  3. Run schema update: python 00-system/skills/notion-master/scripts/manage_database.py update \ --db <database_id> \ --add-property '{"name": "Priority", "type": "select", "options": ["Low", "Medium", "High"]}'
  4. Run discovery to refresh context

Supported Property Types for Creation:

  • title - Primary name field (required, one per database)
  • rich_text - Multi-line text
  • number - Numeric values (with optional precision)
  • select - Single choice from options
  • multi_select - Multiple choices from options
  • date - Date/datetime
  • checkbox - Boolean true/false
  • url, email, phone_number - Validated text fields
  • people - User assignment
  • relation - Link to another database

Workflow 5: Update Page

Triggers: "update [page]", "edit [page]", "change [property] to [value]", "modify page"

Purpose: Modify properties of an existing page.

Steps:

  1. Run config check (Workflow 0)
  2. Identify page:

- By page ID if known - By title search in database: python search_skill_database.py --db <id> --filter "Name contains [search]"

  1. Show current properties
  2. Accept changes from user
  3. Run update: python 00-system/skills/notion-master/scripts/manage_page.py update \ --page <page_id> \ --properties '{"Status": "Done", "Priority": "High"}'
  4. Confirm changes with updated page details

Workflow 6: Get/Delete Page

Triggers: "get page [id]", "show page details", "delete page", "remove [page]"

Purpose: Retrieve full page details or delete a page.

Steps:

Get Page

  1. Run config check
  2. Run retrieval: python 00-system/skills/notion-master/scripts/manage_page.py get --page <page_id>
  3. Display all properties in formatted output

Delete Page (Archive)

  1. Run config check
  2. Confirm with user: "Are you sure you want to delete [page title]?"
  3. Run deletion: python 00-system/skills/notion-master/scripts/manage_page.py delete --page <page_id>
  4. Confirm page archived (Notion doesn't hard-delete, only archives)

Workflow 7: Manage Content/Blocks

Triggers: "append to [page]", "add section to [page]", "edit content of [page]", "list blocks", "delete block"

Purpose: Read, append, update, and delete content blocks within pages.

Steps:

List Page Content

  1. Run config check (Workflow 0)
  2. Get page blocks: python 00-system/skills/notion-master/scripts/manage_blocks.py children --page <page_id>
  3. Display block hierarchy with types and previews

Append Content

  1. Run config check
  2. Identify target page/block
  3. Build block content (simple or complex): # Simple block python 00-system/skills/notion-master/scripts/manage_blocks.py append \ --page <page_id> --type paragraph --text "Your content here" # Multiple blocks python 00-system/skills/notion-master/scripts/manage_blocks.py append \ --page <page_id> --content '[{"type": "heading_1",...}, {"type": "paragraph",...}]'
  4. Confirm appended blocks

Update Block

  1. Run config check
  2. Get block ID (from children list or user)
  3. Update block content: python 00-system/skills/notion-master/scripts/manage_blocks.py update \ --block <block_id> --content '{"paragraph": {"rich_text": [...]}}'
  4. Confirm update

Delete Block

  1. Run config check
  2. Confirm with user: "Delete block [preview]?"
  3. Delete block: python 00-system/skills/notion-master/scripts/manage_blocks.py delete --block <block_id> --confirm

Supported Block Types (load notion-master/references/block-types.md for full list):

  • Text: paragraph, heading_1, heading_2, heading_3, quote, callout
  • Lists: bulleted_list_item, numbered_list_item, to_do, toggle
  • Code: code (with language), equation
  • Media: image, video, file, bookmark
  • Structure: divider, table_of_contents

Workflow 8: User Management

Triggers: "list users", "who can access", "get user", "find user"

Purpose: List workspace users and get user details for @mentions.

Steps:

List All Users

  1. Run config check (Workflow 0)
  2. List workspace users: python 00-system/skills/notion-master/scripts/manage_users.py list --save
  3. Display user list with IDs (for people properties and @mentions)
  4. Optionally save to 01-memory/integrations/notion-users.yaml

Get Current Bot

  1. Run config check
  2. Get bot info: python 00-system/skills/notion-master/scripts/manage_users.py me
  3. Display integration details

Workflow 9: Comments

Triggers: "add comment", "list comments", "comment on [page]", "reply to comment"

Purpose: Create and list comments on pages and discussions.

Steps:

List Comments

  1. Run config check (Workflow 0)
  2. Get page/block ID
  3. List comments: python 00-system/skills/notion-master/scripts/manage_comments.py list --page <page_id>
  4. Display comments with authors and timestamps

Create Comment

  1. Run config check
  2. Identify target page or discussion
  3. Get comment text from user
  4. Create comment: # New comment on page python 00-system/skills/notion-master/scripts/manage_comments.py create \ --page <page_id> --text "Your comment here" # Reply to existing discussion python 00-system/skills/notion-master/scripts/manage_comments.py create \ --discussion <discussion_id> --text "Your reply here"
  5. Confirm comment created

Workflow 10: Advanced Operations (Future)

Status: Planned for Phase 6

Sub-workflows:

  • "upload file to [page]" → File upload flow (3-step process)
  • "create data source" → Data source API (2025 feature)

Context File Format

Location: 01-memory/integrations/notion-databases.yaml

---
last_synced: 2025-12-10T23:00:00
sync_count: 3
databases:
  - id: "abc123-def456"
    name: "Client Projects"
    parent: "Marketing"  # For disambiguation
    url: "https://notion.so/..."
    properties:
      - name: "Name"
        type: "title"
      - name: "Status"
        type: "select"
        options: ["Not Started", "In Progress", "Complete"]
      - name: "Priority"
        type: "select"
        options: ["Low", "Medium", "High"]
      - name: "Due Date"
        type: "date"
      - name: "Assignee"
        type: "people"
  - id: "ghi789-jkl012"
    name: "Content Calendar"
    parent: null
    url: "https://notion.so/..."
    properties:
      - name: "Title"
        type: "title"
      - name: "Publish Date"
        type: "date"
      - name: "Channel"
        type: "multi_select"
        options: ["Blog", "Twitter", "LinkedIn"]
---

# Notion Databases Context

Auto-generated by notion-connect skill.
Run "refresh notion" to update.

Fuzzy Matching Logic

When user says a database name:

  1. Exact match: "Client Projects" → finds "Client Projects"
  2. Case-insensitive: "client projects" → finds "Client Projects"
  3. Partial match: "Projects" → finds "Client Projects", "Personal Projects"
  4. Word match: "calendar" → finds "Content Calendar", "Editorial Calendar"

Disambiguation prompt (if multiple matches):

Found 3 databases matching "Projects":
1. Client Projects (Marketing)
2. Personal Projects
3. Project Templates

Which one? (enter number or be more specific)

Error Handling

ErrorAction
No databases foundCheck Notion integration has database access
Database not in contextRun discovery: "refresh notion"
Permission deniedVerify integration connected to database
Rate limitedWait and retry with exponential backoff
Invalid propertyShow valid options from cached schema

Load notion-master/references/error-handling.md for detailed troubleshooting.


Examples

First-time setup:

User: "connect to notion"
AI: Running Notion discovery...
    Found 5 databases: Client Projects, Content Calendar, CRM, Tasks, Meeting Notes
    Saved to context. You can now query any database by name!

Query by name:

User: "query my Projects for active items"
AI: Querying "Client Projects"...
    Found 12 active items:
    1. Website Redesign (High priority, due Jan 15)
    2. Q1 Marketing Plan (Medium priority, due Jan 30)
    ...

Create page:

User: "add a new task to my Tasks database"
AI: Creating in "Tasks" database.
    Required properties:
    - Task Name: [your input]
    - Status: Not Started / In Progress / Complete
    - Due Date: [optional, format: YYYY-MM-DD]

    What's the task name?

*Notion Connect v1.1 - Meta-skill for complete Notion workspace integration*


Changelog:

  • v1.1 (2025-12-11): Added first-time user setup section, improved exit code documentation
  • v1.0 (2025-12-10): Initial release with 10 workflows

平台分布

Claude Code

29.51%
按下载量换算58

Antigravity

23.07%
按下载量换算45

windsurf

15.29%
按下载量换算30

Codex

10.72%
按下载量换算21

OpenCode

7.38%
按下载量换算14

Gemini CLI

2.92%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。

继续浏览同类 Skills