← 返回 Skills 市场
718
总下载
0
收藏
12
当前安装
3
版本数
在 OpenClaw 中安装
/install pdf-toolkit
功能描述
Run a local script to work with PDF files, DOCX documents, OCR, and text-to-speech. Use the read tool to load this SKILL.md, then exec the uv run command ins...
安全使用建议
This skill appears to do what it says: local PDF/DOCX processing, OCR, and text→speech. Before installing or running it: 1) Be aware it operates on any filesystem path you provide — do not pass paths to secrets or system files you don't want read or written. 2) TTS uses edge-tts (network): avoid sending sensitive text to TTS because it will leave the host. 3) The tool uses 'uv run' to fetch Python packages at runtime (packages are listed but not version-pinned); if you require stricter supply-chain controls, run it in an isolated environment or pin/verify dependency versions. 4) Optional binaries (ffmpeg, tesseract, pdfimages, pandoc, libreoffice) are only needed for their respective features — install them only if you use those features. 5) If you want maximum safety, review the included source files locally and run the skill in a sandboxed container before granting it access to important host paths.
功能分析
Type: OpenClaw Skill
Name: pdf-toolkit
Version: 0.0.3
The pdf-toolkit bundle is a comprehensive utility for PDF and document manipulation, including OCR and text-to-speech capabilities. It uses standard Python libraries (pypdf, reportlab) and system binaries (tesseract, ffmpeg, pandoc) for its stated purposes. While it performs network requests for the TTS feature (via edge-tts) and has broad file system access, these behaviors are explicitly disclosed and warned about in SKILL.md. The code follows security best practices by using list-based arguments in subprocess calls to prevent shell injection.
能力评估
Purpose & Capability
The name/description match the included code: the scripts provide PDF metadata, extraction, merging, splitting, OCR, DOCX read/write, conversion, and TTS. The only required binary declared in the registry is 'uv', which matches the SKILL.md and the included wrapper (pdf-toolkit.sh) that executes 'uv run'. Optional system binaries (ffmpeg, tesseract, pdfimages, pandoc, libreoffice) are documented in SKILL.md and used only when those features are invoked.
Instruction Scope
SKILL.md instructs the agent to run 'uv run' on the included main.py and documents which system tools are optional. The skill explicitly states it will read/write any host path the caller supplies; the code resolves and opens absolute paths, so if a user provides an arbitrary system path the skill will read/write it. The only external network behavior described is edge-tts for TTS. This broad filesystem access is coherent for a file-manipulation tool but is a sensitive capability the user should be aware of.
Install Mechanism
There is no formal install spec; this is run via 'uv run', and main.py includes a header listing Python dependencies. That means uv will fetch and install (or use) packages from the Python ecosystem at runtime. Dependencies are not pinned to specific versions in the header. Runtime installation from PyPI is expected for this skill's Python libraries, but it increases the attack surface compared to a purely local, pre-vetted bundle.
Credentials
The skill requests no environment variables or credentials (primaryEnv none). That's proportionate. However, it will send TTS text to an external service (edge-tts) if TTS is used — SKILL.md warns about this. The code will access any filesystem path supplied by the caller (including ~ and absolute paths), which is necessary for the tool but can expose sensitive files if the user supplies them.
Persistence & Privilege
always:false and user-invocable:true; the skill does not request elevated/system-wide persistence and does not modify other skills or agent-wide configuration. It runs on demand and does not require automatic always-on privileges.
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 pdf-toolkit - 安装完成后，直接呼叫该 Skill 的名称或使用
/pdf-toolkit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
v0.0.3
### v0.0.3 Changelog
- Added "Chat Delivery" section to SKILL.md with guidelines for sending generated files as chat attachments and cleaning up temporary files.
- Clarified file delivery and deletion behavior for output files in chat interfaces.
- No functional code or command changes; documentation update only.
v0.0.2
- Added a new "File Access And Network Behavior" section to document file handling and network usage.
- Clarified that the skill can read from and write to any file path provided by the caller, not just workspace paths.
- Documented that the `tts` command uses `edge-tts` and sends text over the network; warned users not to send sensitive/private text.
- Explained that all other commands run locally, except for required optional binaries.
- No changes to supported commands or core functionality.
v0.0.1
Initial release of pdf-toolkit.
- Provides commands to extract text, tables, and images from PDFs.
- Supports merging, splitting, rotating, and creating PDFs.
- Includes DOCX reading and writing capabilities.
- Adds OCR support for scanned PDFs (requires tesseract).
- Offers text-to-speech output to MP3 (requires ffmpeg).
- Enables document format conversion (requires pandoc or libreoffice).
- `doctor` command checks availability of optional system tools.
元数据
常见问题
Pdf Toolkit 是什么？
Run a local script to work with PDF files, DOCX documents, OCR, and text-to-speech. Use the read tool to load this SKILL.md, then exec the uv run command ins... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 718 次。
如何安装 Pdf Toolkit？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install pdf-toolkit」即可一键安装，无需额外配置。
Pdf Toolkit 是免费的吗？
是的，Pdf Toolkit 完全免费（开源免费），可自由下载、安装和使用。
Pdf Toolkit 支持哪些平台？
Pdf Toolkit 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Pdf Toolkit？
由 youpele52（@youpele52）开发并维护，当前版本 v0.0.3。
推荐 Skills