Token导航 LogoToken导航

Agent Skill · 研究检索

paperless-ngx(无纸化 NGX)

paperless-ngx 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。

paperless-ngx 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;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,674 个 Skill 里排第 71,089。

总安装 196周安装 8GitHub Stars 35下载 63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
unknown

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:paperless-ngx(无纸化 NGX)
来源仓库:https://github.com/jmagar/claude-homelab
仓库路径:skills/paperless-ngx
安装命令:
npx skills add https://github.com/jmagar/claude-homelab --skill paperless-ngx
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/jmagar/claude-homelab --skill paperless-ngx

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。

paperless-ngx 技能聚焦于文档管理与检索支持。可用于构建私有知识库或自动化归档流程。paperless-ngx 属于研究检索类 Skill,可作为该场景下的辅助能力补充。安装后可在 Claude 等平台调用,需配合本地实例使用。若涉及文件读写,务必确保路径安全与权限合规。

建议先行部署 Paperless-NGX 服务端再启用此技能。

来源仓库内路径:skills/paperless-ngx

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/jmagar/claude-homelab --skill paperless-ngx;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

paperless-ngx 怎么安装?

在终端执行 npx skills add https://github.com/jmagar/claude-homelab --skill paperless-ngx。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

paperless-ngx 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。

paperless-ngx 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。

paperless-ngx 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

paperless-ngx 用的人多吗?

公开数据是累计安装 196 次,近一周安装 8 次,下载 63 次,GitHub 35 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。

paperless-ngx 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

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。

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。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。

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。
  • 需要明确授权才能用到公司项目里的场景。
  • 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
paperless-ngx

无纸化 NGX · 本页

19635接触敏感数据未明确有提醒

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

Paperless-ngx Skill

⚠️ MANDATORY SKILL INVOCATION ⚠️

YOU MUST invoke this skill (NOT optional) when the user mentions ANY of these triggers:

  • "upload document", "add document to paperless", "scan to paperless"
  • "search paperless", "find document", "search documents"
  • "tag document", "add tags", "manage tags in paperless"
  • "manage correspondents", "add correspondent", "list correspondents"
  • "update document", "edit document metadata", "change document type"
  • "archive document", "export document", "download document"
  • "delete document", "remove document from paperless"
  • "list my documents", "show recent documents", "what documents do I have"
  • Any mention of Paperless-ngx, document management, OCR, or paperless office

Failure to invoke this skill when triggers occur violates your operational requirements.


Purpose

This skill provides read-write access to a self-hosted Paperless-ngx instance for document management with OCR. Paperless-ngx transforms physical documents into a searchable online archive with full-text search, tagging, and metadata management.

Core capabilities:

  • Upload documents with auto-OCR and metadata extraction
  • Search documents by content, tags, correspondent, or date
  • Manage tags, correspondents, and document types
  • Update document metadata (title, tags, correspondent, dates)
  • Archive and export operations
  • Delete documents (with confirmation)
  • Bulk operations on multiple documents

Primary use case: Maintain a searchable digital archive of all documents with powerful organization and retrieval capabilities.

Setup

Prerequisites

  • Paperless-ngx instance running and accessible
  • API token generated from Paperless-ngx UI
  • curl and jq installed

Credential Configuration

Add these variables to ~/.claude-homelab/.env:

# Paperless-ngx - Document management system
PAPERLESS_URL="https://paperless.example.com"
PAPERLESS_API_TOKEN="<your_api_token>"

To generate an API token:

  1. Log into your Paperless-ngx instance
  2. Go to Settings → My Profile
  3. Click "Create Token" under "API Tokens"
  4. Copy the generated token
  5. Add to .env file as shown above

Security:

  • .env file is gitignored (never commit)
  • Set permissions: chmod 600 ~/.claude-homelab/.env
  • Token has same permissions as your user account

Commands

All commands return JSON output for LLM parsing. Scripts source credentials from .env automatically.

Document Operations

Upload a document:

bash scripts/paperless-api.sh upload /path/to/document.pdf
bash scripts/paperless-api.sh upload scan.jpg --title "Receipt" --tags "expense,2024"
bash scripts/paperless-api.sh upload contract.pdf --correspondent "Acme Corp" --document-type "Contract"

Search documents:

bash scripts/paperless-api.sh search "invoice"
bash scripts/paperless-api.sh search "meeting notes" --limit 10
bash scripts/paperless-api.sh search "2024" --tags "tax"
bash scripts/paperless-api.sh search --correspondent "John Doe"

List documents:

bash scripts/paperless-api.sh list
bash scripts/paperless-api.sh list --limit 20
bash scripts/paperless-api.sh list --ordering "-created"

Get document details:

bash scripts/paperless-api.sh get <document-id>

Download/export document:

bash scripts/paperless-api.sh download <document-id>
bash scripts/paperless-api.sh download <document-id> --output /path/to/save.pdf

Update document:

bash scripts/paperless-api.sh update <document-id> --title "New Title"
bash scripts/paperless-api.sh update <document-id> --add-tags "urgent,reviewed"
bash scripts/paperless-api.sh update <document-id> --correspondent "Jane Smith"
bash scripts/paperless-api.sh update <document-id> --document-type "Invoice"
bash scripts/paperless-api.sh update <document-id> --archive-serial-number "2024-001"

Delete document:

bash scripts/paperless-api.sh delete <document-id>  # Prompts for confirmation

Tag Management

List tags:

bash scripts/tag-api.sh list
bash scripts/tag-api.sh list --ordering "name"

Create tag:

bash scripts/tag-api.sh create "project-alpha"
bash scripts/tag-api.sh create "urgent" --color "#ff0000"

Get tag details:

bash scripts/tag-api.sh get <tag-id>

Update tag:

bash scripts/tag-api.sh update <tag-id> --name "new-name"
bash scripts/tag-api.sh update <tag-id> --color "#00ff00"

Delete tag:

bash scripts/tag-api.sh delete <tag-id>  # Prompts for confirmation

Correspondent Management

List correspondents:

bash scripts/correspondent-api.sh list

Create correspondent:

bash scripts/correspondent-api.sh create "Acme Corporation"

Get correspondent details:

bash scripts/correspondent-api.sh get <correspondent-id>

Update correspondent:

bash scripts/correspondent-api.sh update <correspondent-id> --name "New Name"

Delete correspondent:

bash scripts/correspondent-api.sh delete <correspondent-id>  # Prompts for confirmation

Bulk Operations

Bulk tag documents:

bash scripts/bulk-api.sh add-tag <tag-id> --documents "1,2,3"
bash scripts/bulk-api.sh remove-tag <tag-id> --documents "1,2,3"

Bulk set correspondent:

bash scripts/bulk-api.sh set-correspondent <correspondent-id> --documents "1,2,3"

Bulk set document type:

bash scripts/bulk-api.sh set-document-type <type-id> --documents "1,2,3"

Bulk delete documents:

bash scripts/bulk-api.sh delete --documents "1,2,3"  # Prompts for confirmation

Workflow

When the user asks about Paperless-ngx:

  1. "Upload this document to Paperless"

- Save file locally if needed - Use upload command with appropriate metadata - Optionally add tags, correspondent, document type - Paperless will auto-OCR and extract text

  1. "Find my 2024 tax documents"

- Use search "tax" --tags "2024" or similar query - Present results with ID, title, date, tags - User can request full details with get <id>

  1. "Tag all invoices from Acme Corp as paid"

- Search for documents: search "invoice" --correspondent "Acme Corp" - Create or find "paid" tag - Use bulk operation to add tag to results

  1. "Show me documents without tags"

- Use list with filter parameter - Present untagged documents - Suggest tagging workflow

  1. "Delete this old document"

- Confirm document ID with user - Ask for explicit confirmation - Use delete <id> command

  1. "Add a new supplier as correspondent"

- Use correspondent-api.sh create "Supplier Name" - Return correspondent ID for future document uploads

Detailed Flow: Document Upload

User: "Upload this receipt to Paperless and tag it as expense"

1. Verify file path exists and is readable
2. Upload document with metadata:
   bash scripts/paperless-api.sh upload /path/to/receipt.pdf --tags "expense"
3. Paperless processes document (OCR, thumbnail generation)
4. Return document ID and success confirmation
5. Optionally ask if user wants to set correspondent or document type

Detailed Flow: Search and Organize

User: "Find all documents from last month that need review"

1. Calculate date range (last month)
2. Search documents by date range
3. Filter results by tag or keyword "review"
4. Present results with metadata
5. Offer to bulk-tag results or export list

Detailed Flow: Bulk Tagging

User: "Tag all invoices from Q1 as archived"

1. Search for invoices in Q1 date range
2. Extract document IDs from results
3. Find or create "archived" tag
4. Use bulk operation to add tag to all documents
5. Report number of documents tagged

Notes

API Details

  • Authentication: Token authentication via Authorization: Token <token> header
  • Base URL: /api/ endpoint
  • API Version: Version 5 (specify via Accept header)
  • Rate limits: No documented limits (self-hosted)
  • Pagination: Uses page and page_size parameters

Document Processing

Paperless-ngx automatically processes uploaded documents:

  • OCR extraction (if not already text-based PDF)
  • Thumbnail generation
  • Metadata extraction (date, correspondent guessing)
  • Full-text indexing for search

Search Syntax

Search supports multiple query types:

  • Full-text: search "meeting notes"
  • Tag filter: search --tags "work,urgent"
  • Correspondent: search --correspondent "John Doe"
  • Date range: search --created-after "2024-01-01"
  • Document type: search --document-type "Invoice"

Tags vs Correspondents

  • Tags: General-purpose labels (e.g., "urgent", "personal", "2024")
  • Correspondents: People or organizations (e.g., "Acme Corp", "John Smith")
  • Document Types: Categories of documents (e.g., "Invoice", "Contract", "Receipt")

All three provide different organizational axes for your documents.

Destructive Operations

Delete operations require confirmation:

  • Delete document: Permanently removes document and files
  • Delete tag: Removes tag from all documents (documents remain)
  • Delete correspondent: Removes from all documents (documents remain)

Always confirm with user before executing delete operations.

Best Practices

  1. Tag consistently: Use lowercase, hyphens for multi-word (e.g., "project-alpha")
  2. Set correspondents: Makes searching by sender/recipient easier
  3. Use document types: Categories help with organization and workflows
  4. Archive serial numbers: Track original paper document storage
  5. Search before upload: Avoid duplicates

Common Errors

401 Unauthorized:

  • Check API token in .env
  • Token may have expired (regenerate in Paperless UI)
  • Verify user permissions

404 Not Found:

  • Verify document/tag/correspondent ID exists
  • Check PAPERLESS_URL is correct

400 Bad Request:

  • Invalid metadata (e.g., non-existent tag ID)
  • Malformed file upload
  • Check script output for specific error message

Connection refused:

  • Paperless-ngx instance not running
  • Verify URL in .env

Reference

  • Official Docs: https://docs.paperless-ngx.com/
  • API Reference: See references/api-endpoints.md for complete API documentation
  • Quick Reference: See references/quick-reference.md for command examples
  • Troubleshooting: See references/troubleshooting.md for common issues
  • Scripts: skills/paperless-ngx/scripts/

🔧 Agent Tool Usage Requirements

CRITICAL: When invoking scripts from this skill via the zsh-tool, ALWAYS use pty: true.

Without PTY mode, command output will not be visible even though commands execute successfully.

Correct invocation pattern:

<invoke name="mcp__plugin_zsh-tool_zsh-tool__zsh">
<parameter name="command">./skills/paperless-ngx/scripts/paperless-api.sh [args]</parameter>
<parameter name="pty">true</parameter>
</invoke>

平台分布

Codex

34.55%
按下载量换算22

Claude

29.46%
按下载量换算19

Cursor

19.62%
按下载量换算12

Gemini CLI

9.86%
按下载量换算6

安全审计

Gen Agent Trust Hub

可疑

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills