← 返回 Skills 市场

PDF文档处理工具

Name: PDF文档处理工具
Author: freedompixels

作者 freedompixels · GitHub ↗ · v1.1.0 · MIT-0

cross-platform ✓ 安全检测通过

总下载

当前安装

版本数

在 OpenClaw 中安装

/install cn-pdf-assistant

功能描述

PDF文档处理工具。本地处理PDF文件，支持文本提取、智能摘要、表格导出、关键词问答、PDF拆分。纯本地处理，保护文档隐私。

使用说明 (SKILL.md)

功能

PDF文本提取（支持指定页码范围）
智能摘要生成（章节标题识别+关键词频率分析）
表格提取（pdfplumber引擎）
关键词问答（基于段落匹配）
PDF按页拆分
纯本地处理，无需联网

使用方法

python3 scripts/pdf_assistant.py \x3CPDF文件路径> --action text
python3 scripts/pdf_assistant.py \x3CPDF文件路径> --action summary
python3 scripts/pdf_assistant.py \x3CPDF文件路径> --action tables
python3 scripts/pdf_assistant.py \x3CPDF文件路径> --action ask --question "关键词"
python3 scripts/pdf_assistant.py \x3CPDF文件路径> --action split

依赖

Python 3.7+
PyPDF2, pdfplumber, pandas, openpyxl

权限声明

读取本地PDF文件
生成输出文件

使用场景

论文阅读：快速提取核心内容
合同审查：提取关键条款
财报分析：提取表格数据
资料整理：批量拆分PDF文档

安全使用建议

This skill appears to do what it claims: local PDF text/table extraction, simple summaries, Q&A, and splitting. Notes before installing: (1) It will create files when you use the split action (default ~/Documents/PDFAssistant/split) — expect local disk writes. (2) SKILL.md lists pandas/openpyxl for table export, but the provided script does not actually export to Excel; if you need that feature verify or extend the script. (3) Ensure required Python packages (PyPDF2, pdfplumber, etc.) are installed from trusted sources and run the script offline if you want to guarantee no network activity. If you want stronger assurance, review/run the script in a controlled environment (sandbox) before using it on sensitive documents.

功能分析

Type: OpenClaw Skill Name: cn-pdf-assistant Version: 1.1.0 The skill bundle is a legitimate PDF processing utility that performs text extraction, summarization, table parsing, and file splitting. The Python script (scripts/pdf_assistant.py) uses standard libraries like pdfplumber and PyPDF2, operates entirely locally, and contains no network calls, obfuscation, or suspicious file system operations. All functionalities align with the documentation in SKILL.md.

能力评估

ℹ Purpose & Capability

Name/description (local PDF processing: text, summary, tables, Q&A, split) match the included script which implements those actions locally. Minor mismatch: SKILL.md lists pandas/openpyxl and '表格导出' but the script extracts table data but does not implement exporting to Excel (pandas/openpyxl are not used). The script docstring mentions translation though no translation code is present.

✓ Instruction Scope

SKILL.md instructs running the included script on a local PDF and the script only reads the provided PDF and writes output files (split files) or prints results. There are no instructions to read unrelated files, environment variables, or to send data externally.

✓ Install Mechanism

No install spec (instruction-only skill with a bundled script). This is low-risk: nothing is downloaded or executed beyond the provided Python script. Dependencies are standard Python packages available on PyPI.

✓ Credentials

The skill requires no environment variables, no credentials, and no config paths. The script only operates on the user-supplied PDF and writes outputs to either the specified output directory or a default ~/Documents/PDFAssistant/split. No secret access is requested.

✓ Persistence & Privilege

The skill does not request always:true and does not modify other skills or system-wide settings. It creates local output directories/files when splitting PDFs (standard, scoped filesystem writes).

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install cn-pdf-assistant
安装完成后，直接呼叫该 Skill 的名称或使用 /cn-pdf-assistant 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.2.0

- Added explicit version field ("1.2.0") to the skill metadata. - No changes to features or usage instructions.

v1.1.0

v1.1.0: fix SKILL.md description format, remove unsupported OCR claims

v1.0.0

cn-pdf-assistant 1.0.0 - 初始发布，支持本地PDF文档处理，保护隐私，无需上传。 - 提供内容问答、智能摘要、表格提取、中英文翻译、页面操作等核心功能。 - 适用于论文阅读、合同审查、财报分析及资料整理等多种场景。 - 支持文本提取、智能摘要、表格结构化导出、关键词定位问答、PDF拆分。 - 可选支持OCR扫描件识别及高级问答功能。

元数据

Slug cn-pdf-assistant

版本 1.1.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 2

常见问题