AI PDF Converter

AI-powered PDF converter using MinerU API. Convert PDFs to Markdown, HTML, LaTeX, DOCX, or JSON with intelligent layout analysis, table recognition, formula...

Security Scan

The skill's stated purpose (MinerU-based PDF conversion) mostly matches its instructions, but it omits credential and privacy guidance and relies on an unverified npm CLI—this gap and the implicit uploading of user PDFs to a remote service are concerning.

Purpose & Capability

Name and description match the SKILL.md instructions: it instructs use of a mineru-open-api CLI to convert PDFs into multiple formats. Requiring the mineru-open-api tool is coherent with the claimed purpose. Minor note: the skill does not declare any credentials or config even though advanced 'extract' actions likely require MinerU API access.

Instruction Scope

Instructions tell the agent to run mineru-open-api commands on local PDF files and to write outputs to ~/MinerU-Skill/<name>_<hash>/. They imply sending documents to MinerU (remote service) but do not explicitly state that files are uploaded, how long they are retained, or how to provide/secure API credentials. For sensitive documents, automatic upload to an external API is a privacy risk and should be called out. The workflow also recommends npm -g installation without guidance on sandboxing or verifying the package source.

Install Mechanism

There is no formal install spec in the registry (instruction-only), but SKILL.md tells users to run npm install -g mineru-open-api. Asking users to install an npm package is reasonable for a CLI-based skill, but the package and its publisher are not referenced (no homepage or source). Installing an unverified global npm package is a moderate risk; the skill does not provide provenance or verification steps.

Credentials

Registry metadata declares no required env vars or primary credential, yet SKILL.md distinguishes between 'flash-extract (no token)' and other extract modes (precision, VLM) that likely require an API token or key. The absence of any declared credential is a mismatch. The skill may require MinerU credentials or other config (API key, endpoint) but gives no instructions for how to set or protect them.

Persistence & Privilege

The skill does not request persistent presence (always is false) and does not modify other skills or system settings. It only writes outputs to a directory under the user's home by default, which is a typical behavior for a conversion tool.

What to consider before installing

Before installing or using this skill: (1) Expect that conversion (except 'flash-extract') will upload PDFs to MinerU servers—do not use it on sensitive documents unless you accept that and have checked MinerU's privacy/retention policy. (2) Verify the npm package 'mineru-open-api' on the npm registry: check the publisher, repository, and recent activity; inspect package source before a global install. (3) Ask the skill author (or vendor site) whether and how API keys are required and where they should be stored (env vars, config file); do not pass secrets into an unverified tool. (4) Prefer running the CLI in a sandbox or disposable container and monitor network traffic if you must test it. If you want, provide the mineru-open-api package URL or the MinerU docs and I can re-evaluate with that context.

AI PDF Converter with mineru-open-api

You are an AI PDF conversion specialist. Convert PDFs to any format using mineru-open-api.

Installation

npm install -g mineru-open-api

Conversion Workflow

1. Quick Markdown (no token):

   mineru-open-api flash-extract document.pdf -o ./output/
   

2. Multi-format conversion:

   mineru-open-api extract document.pdf -f md,html,latex,docx -o ./output/
   

3. AI-enhanced conversion (VLM):

   mineru-open-api extract complex.pdf -f html --model vlm -o ./output/
   

4. Batch conversion:

   mineru-open-api extract *.pdf -f html -o ./results/
   

Output Formats

Format	flash-extract	extract
Markdown	Yes	Yes
HTML	No	Yes
LaTeX	No	Yes
DOCX	No	Yes
JSON	No	Yes

Key Rules

• Default to flash-extract for simple Markdown conversion under 10MB/20 pages
• Use extract for HTML, LaTeX, DOCX, JSON, or complex documents
• --model vlm for best accuracy on complex layouts
• --model pipeline for reliability (no hallucination)
• Generate default output dir: ~/MinerU-Skill/<name>_<hash>/