Doc Parse

Parse and extract structured content from Word documents (.doc, .docx) into well-organized Markdown using MinerU. Preserves the full document hierarchy: head...

Security Scan

The skill mostly matches its stated purpose (using the mineru-open-api CLI to parse Word files), but there is an inconsistency around required credentials and a few minor provenance/installation considerations the user should verify before installing.

Purpose & Capability

Name and description match the declared requirements: the skill uses the mineru-open-api CLI to parse .doc/.docx into Markdown. Requesting the mineru-open-api binary and a MINERU_TOKEN is consistent with a cloud-backed CLI. However, SKILL.md explicitly says 'flash-extract' for .docx requires no token while the registry metadata lists MINERU_TOKEN as a required env var/primary credential — this is an internal inconsistency.

Instruction Scope

SKILL.md instructs only running the mineru-open-api CLI on local files or URLs and how to set MINERU_TOKEN. It does not direct reading unrelated system files, scanning shell history, or exfiltrating data to unexpected endpoints. The instructions do not clarify whether processing happens locally or via MinerU's API; that ambiguity is important for privacy but not a scope creep in itself.

Install Mechanism

Install options are standard: an npm package 'mineru-open-api' and a Go 'github.com/opendatalab/...' install path (GitHub). These are expected for a CLI authored by the MinerU project. No arbitrary binary downloads or obscure URLs are present in the install spec.

Credentials

Requesting MINERU_TOKEN as the primary credential is reasonable for an API-backed CLI and required for some operations (.doc extract). But marking MINERU_TOKEN as globally required conflicts with SKILL.md which says quick .docx 'flash-extract' requires no token. Also the skill does not declare additional unrelated credentials, which is good.

Persistence & Privilege

always:false (default) and autonomous invocation allowed (platform default). The skill does not request persistent system-wide privileges or modification of other skills/configuration in the provided instructions.

What to consider before installing

This skill appears to do what it says (wrap the MinerU CLI to parse .doc/.docx), but there are two things to check before installing: (1) credential discrepancy — the registry metadata marks MINERU_TOKEN as required, yet SKILL.md says flash-extract for .docx works without a token; verify whether you can use quick parsing without providing a token. (2) data handling/privacy — the CLI likely communicates with MinerU's service (mineru.net) for at least some operations; confirm whether document content is uploaded, how long it's retained, and whether that is acceptable for your documents. Also confirm the npm package and GitHub repo are the official MinerU releases (inspect the mineru-open-api source) and only provide a token with minimal necessary scope. If you need higher assurance, run the CLI in a sandbox or review the open-source repo and network calls before sending sensitive documents.

Parse Word (.doc/.docx) documents into structured Markdown using MinerU. Preserves document hierarchy including headings, lists, tables, and paragraphs.

Install

npm install -g mineru-open-api
# or via Go (macOS/Linux):
go install github.com/opendatalab/MinerU-Ecosystem/cli/mineru-open-api@latest

Quick Start

# Quick parse from .docx (no token required)
mineru-open-api flash-extract report.docx

# Save structured Markdown to directory
mineru-open-api flash-extract report.docx -o ./out/

# Parse .doc file (requires token)
mineru-open-api extract report.doc -o ./out/

# With language hint
mineru-open-api extract report.docx --language en -o ./out/

Authentication

No token needed for flash-extract on .docx. Token required for .doc:

mineru-open-api auth             # Interactive token setup
export MINERU_TOKEN="your-token" # Or via environment variable

Create token at: https://mineru.net/apiManage/token

Capabilities

• Supported input: .doc, .docx (local file or URL)
• .docx: supports flash-extract (no token, max 10 MB / 20 pages) and extract
• .doc: requires extract with token
• Output preserves document structure as Markdown hierarchy
• Language hint with --language (default: ch, use en for English)

Notes

• .doc requires extract with token; .docx supports flash-extract for quick parsing
• Output goes to stdout by default; use -o <dir> to save to a file or directory
• All progress/status messages go to stderr; document content goes to stdout
• MinerU is open-source by OpenDataLab (Shanghai AI Lab): https://github.com/opendatalab/MinerU