← Back to Skills Marketplace
113
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install pdflux-pdf2markdown
Description
Convert unstructured documents into LLM-ready structured data. Supports PDF, Word, PPT, and images; extracts paragraphs, formulas, tables, charts, and other...
Usage Guidance
This skill will upload your local document to a remote service (default: platform.paodingai.com) and return Markdown. Before installing or running it: (1) confirm which environment variable name the runtime expects — the registry lists PAODINGAI_* but the script uses PD_ROUTER_API_KEY/PD_ROUTER_BASE_URL; set the correct variable to avoid accidentally exposing other credentials; (2) verify you trust the remote service and that sending the document is acceptable (contains PII/IP/secret data?); (3) if you need source provenance, ask for the skill's homepage or repository or run the included script in a sandbox; (4) ensure your Node runtime supports fetch/Blob/FormData or test locally; (5) if you need offline conversion or don't want external uploads, do not use this skill. If you decide to proceed, provide only a dedicated API key with minimal scope and test with non-sensitive documents first.
Capability Analysis
Type: OpenClaw Skill
Name: pdflux-pdf2markdown
Version: 1.0.1
The skill is a legitimate tool for converting documents (PDF, Word, etc.) to Markdown using the PaodingAI SaaS platform. The core logic in `scripts/upload_to_markdown.js` uses standard Node.js APIs to upload files, poll for status, and download results from `platform.paodingai.com`. While the skill transmits local files to a remote service, this is its primary stated purpose, and there is no evidence of unauthorized data exfiltration, credential theft, or malicious command execution.
Capability Assessment
Purpose & Capability
The skill's name/description (document -> Markdown conversion) align with the included script which uploads a local file to a remote parsing service and downloads Markdown. However, registry metadata and SKILL.md header declare environment variables named PAODINGAI_API_KEY and PAODINGAI_API_BASE_URL while the runtime instructions and script require PD_ROUTER_API_KEY (and optionally PD_ROUTER_BASE_URL / PD_ROUTER_SERVICE_CODE). This naming mismatch is an incoherence that could cause misconfiguration or confusion.
Instruction Scope
The SKILL.md and bundled script instruct the agent to upload the user's local file to a remote API (default https://platform.paodingai.com/) then poll and download Markdown. That is within the stated purpose, but it is an explicit data exfiltration step: local files are sent off-host. The instructions also insist the bundled script be executed (don't reimplement). There is no unexpected access to unrelated system paths or extra env vars in the script, but the requirement to upload files to an external service is a privacy/security consideration users must accept.
Install Mechanism
There is no remote install or arbitrary download. The skill is instruction-only with a small included Node script. No external archives or installers are fetched by the skill itself.
Credentials
The script needs a single gateway API key to talk to the parsing service (PD_ROUTER_API_KEY). That is proportionate to the task. However the registry metadata lists different env var names (PAODINGAI_API_KEY, PAODINGAI_API_BASE_URL) than the script/README (PD_ROUTER_API_KEY / PD_ROUTER_BASE_URL). This inconsistency increases risk of leaking or mis-supplying credentials, and the skill's source is 'unknown' with no homepage, so the provenance of the API endpoint and key handling cannot be easily verified.
Persistence & Privilege
The skill does not request permanent/always-on inclusion (always:false), and it does not modify other skills or system-wide settings. Normal autonomous invocation is allowed (disable-model-invocation:false) which is expected for user-invoked skills; nothing here grants excessive platform privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install pdflux-pdf2markdown - After installation, invoke the skill by name or use
/pdflux-pdf2markdown - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- Changelog summary: Added metadata and translated documentation to English for broader accessibility.
- Added a metadata section specifying author, version, emoji, required environment variables, and dependencies.
- Translated all documentation and usage instructions from Chinese to English.
- Updated description and behavioral guidelines to reflect the new language and metadata requirements.
- No changes to code or core functionality.
v1.0.0
init
Metadata
Frequently Asked Questions
What is PDFlux-PDF2Markdown?
Convert unstructured documents into LLM-ready structured data. Supports PDF, Word, PPT, and images; extracts paragraphs, formulas, tables, charts, and other... It is an AI Agent Skill for Claude Code / OpenClaw, with 113 downloads so far.
How do I install PDFlux-PDF2Markdown?
Run "/install pdflux-pdf2markdown" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is PDFlux-PDF2Markdown free?
Yes, PDFlux-PDF2Markdown is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does PDFlux-PDF2Markdown support?
PDFlux-PDF2Markdown is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created PDFlux-PDF2Markdown?
It is built and maintained by cumtyc (@cumtyc); the current version is v1.0.1.
More Skills