Doc Genius

支持PDF、Word、Markdown智能摘要和格式转换，提供批量处理与进度报告，提升文档处理效率。

Security Scan

The skill largely matches its document-processing purpose, but contains undocumented network/billing logic and a hard-coded billing API key that do not align with the published description or metadata — review before use.

Purpose & Capability

The name/description and most code files implement PDF/Word/Markdown summarization and conversion, which is coherent. However, a bundled 'paid' variant (scripts/doc_processor_paid.py) includes SkillPay billing integration (skillpay.me) and an embedded BILLING_API_KEY and SKILL_ID that are not mentioned in SKILL.md or registry metadata. The presence of billing code in a tool advertised as a free document processor is unexpected and should be justified.

Instruction Scope

SKILL.md instructs running scripts/doc_processor.py and references OPENAI_API_KEY for AI summarization. It does not mention the paid script or any billing/remote calls. The codebase includes additional scripts (doc_processor_paid.py and v2) that import 'requests' and contact external endpoints; this expands runtime actions beyond the documented instructions and the user-visible examples.

Install Mechanism

No install spec downloads arbitrary code; this is an instruction-and-source bundle. Dependencies are local Python packages (PyPDF2, python-docx, markdown, beautifulsoup4) and no external installers or archive downloads are used.

Credentials

Registry metadata declares no required environment variables, but SKILL.md and code reference OPENAI_API_KEY for AI summarization (expected). More concerning: scripts/doc_processor_paid.py hard-codes a BILLING_API_KEY and a user-specific VENV_PYTHON path. A billing API key embedded in code is disproportionate and sensitive; the skill also performs network calls to billing endpoints without documenting them in metadata or instructions.

Persistence & Privilege

The skill does not request always:true, does not claim to modify other skills, and appears to run as user-invoked scripts. No elevated persistence or automatic always-on behavior is present in the metadata.

Scan Findings in Context

[hardcoded-credential] unexpected: scripts/doc_processor_paid.py contains a hard-coded BILLING_API_KEY ('sk_0de94ea93e9...') and SKILL_ID. A document-processing skill should not include embedded billing credentials in source; this is sensitive and unexpected.

[external-billing-endpoint] unexpected: scripts/doc_processor_paid.py calls https://skillpay.me API endpoints (balance, charge, payment-link). The public SKILL.md and registry metadata do not document billing integration or require network access for payments, so these network calls are not clearly expected.

[user-path-in-source] unexpected: scripts/doc_processor_paid.py contains VENV_PYTHON = '/Users/gaolei/.openclaw/workspace/.venv/bin/python3', a user-specific path. This looks like leaked local configuration and is not relevant to the documented capabilities.

What to consider before installing

This skill's core document-processing code appears legitimate, but there are important red flags you should address before installing or running it: (1) The package includes a 'paid' script that contacts an external billing service (skillpay.me) and contains a hard-coded billing API key — treat that key as sensitive and avoid running that script until you confirm its legitimacy. (2) The registry metadata does not declare environment variables (OPENAI_API_KEY) referenced in the docs; expect to provide your OpenAI key if you plan to use AI summarization. (3) Run the code in a restricted environment (container or sandbox) and inspect or remove the paid script if you do not intend to use billing. (4) Ask the author for provenance: where the hard-coded billing key came from, why billing is bundled but undocumented, and for a version without embedded secrets. If you cannot verify the source or the billing integration, do not run the paid script and consider rejecting this skill.

Doc Genius - 智能文档处理助手

快速开始

智能摘要

# PDF摘要
python3 scripts/doc_processor.py summarize /path/to/document.pdf

# Word摘要
python3 scripts/doc_processor.py summarize /path/to/document.docx

# Markdown摘要
python3 scripts/doc_processor.py summarize /path/to/document.md --format json

格式转换

# PDF → Markdown
python3 scripts/doc_processor.py convert /path/to/document.pdf --output markdown

# Word → Markdown
python3 scripts/doc_processor.py convert /path/to/document.docx --output markdown

# Markdown → HTML
python3 scripts/doc_processor.py convert /path/to/document.md --output html

批量处理

# 批量转换文件夹
python3 scripts/doc_processor.py batch /path/to/folder --output markdown

# 批量摘要
python3 scripts/doc_processor.py batch /path/to/folder --action summarize

---

输出格式

JSON格式（默认）

{
  "file": "document.pdf",
  "type": "pdf",
  "summary": "这是文档的智能摘要...",
  "keywords": ["关键词1", "关键词2"],
  "word_count": 5000,
  "pages": 12
}

Markdown格式

python3 scripts/doc_processor.py summarize document.pdf --format markdown

---

核心功能

1. 智能摘要

支持格式：

• ✅ PDF（PyPDF2）
• ✅ Word（.docx）
• ✅ Markdown
• ✅ 纯文本

摘要算法：

• 本地摘要（TextRank，速度快）
• AI摘要（OpenAI API，质量高）

示例：

# 本地摘要
python3 scripts/doc_processor.py summarize document.pdf --method local

# AI摘要（需配置API Key）
export OPENAI_API_KEY="sk-xxx"
python3 scripts/doc_processor.py summarize document.pdf --method ai

---

2. 格式转换

转换矩阵：

输入格式	输出格式	状态
PDF	Markdown	✅
PDF	HTML	⚠️ 实验性
Word	Markdown	✅
Word	HTML	✅
Markdown	HTML	✅
Markdown	Word	🔜 计划中

示例：

# PDF → Markdown（推荐）
python3 scripts/doc_processor.py convert report.pdf --output markdown

# Word → HTML
python3 scripts/doc_processor.py convert report.docx --output html

---

3. 批量处理

功能：

• 文件夹扫描
• 并发处理
• 进度报告
• 错误日志

示例：

# 批量转换（默认并发数=5）
python3 scripts/doc_processor.py batch /path/to/docs --output markdown

# 指定并发数
python3 scripts/doc_processor.py batch /path/to/docs --output markdown --workers 10

# 生成报告
python3 scripts/doc_processor.py batch /path/to/docs --action summarize --report report.json

---

4. 结构化提取（实验性）

提取内容：

• 标题层级
• 目录
• 关键信息（日期、金额、人名）

示例：

python3 scripts/doc_processor.py extract document.pdf --fields title,toc,dates

---

高级用法

使用AI摘要

# 配置API Key
export OPENAI_API_KEY="sk-xxx"

# AI摘要（更智能）
python3 scripts/doc_processor.py summarize document.pdf --method ai --model gpt-4

自定义输出

# 指定输出文件
python3 scripts/doc_processor.py convert document.pdf --output markdown --out-file output.md

# 指定输出目录
python3 scripts/doc_processor.py batch /path/to/docs --output-dir /path/to/output

过滤处理

# 只处理PDF文件
python3 scripts/doc_processor.py batch /path/to/docs --filter "*.pdf"

# 排除文件
python3 scripts/doc_processor.py batch /path/to/docs --exclude "temp_*"

---

技术细节

依赖库

PyPDF2==3.0.1          # PDF处理
python-docx==1.1.0     # Word处理
markdown==3.5.1        # Markdown处理
beautifulsoup4==4.12.2 # HTML解析
aiofiles==23.2.1       # 异步文件处理

安装依赖

pip install PyPDF2 python-docx markdown beautifulsoup4 aiofiles

---

性能优化

并发处理

• 默认并发数：5
• 最大并发数：20
• 推荐：根据CPU核心数调整

内存优化

• 流式处理大文件（>10MB）
• 分块处理（避免内存溢出）

---

错误处理

常见错误

错误	原因	解决方案
FileNotFoundError	文件不存在	检查路径
PermissionError	权限不足	检查文件权限
UnsupportedFormat	格式不支持	查看支持列表

日志级别

# 调试模式
python3 scripts/doc_processor.py summarize document.pdf --log-level debug

---

最佳实践

1. 大文件处理

# 分块处理
python3 scripts/doc_processor.py summarize large.pdf --chunk-size 1000

2. 批量处理优化

# 使用适当的并发数
python3 scripts/doc_processor.py batch /path/to/docs --workers $(nproc)

3. 输出格式选择

场景	推荐格式
内容分析	JSON
人类阅读	Markdown
网页展示	HTML

---

使用场景

1. 研究人员

• 快速阅读大量论文
• 提取关键信息
• 生成文献摘要

2. 内容创作者

• 转换格式（PDF→Markdown）
• 提取素材
• 智能摘要

3. 企业用户

• 批量处理合同
• 文档格式统一
• 知识库构建

---

与其他技能配合

scrapling-fetch

# 抓取网页 → 转换PDF → 智能摘要
python3 scrapling-fetch/scripts/fetch.py "https://example.com/article" --text > temp.md
python3 doc-genius/scripts/doc_processor.py summarize temp.md

---

更新日志

v1.0.0 (2026-03-07)

• ✅ 初始发布
• ✅ 支持PDF/Word/Markdown摘要
• ✅ 支持格式转换
• ✅ 支持批量处理

---

反馈与支持

• GitHub Issues: [待补充]
• ClawHub: https://clawhub.com/skill/doc-genius
• Email: [待补充]

---

Doc Genius - 让文档处理更智能 📄✨