HTML OCR

OCR for HTML pages containing image-embedded or scanned content. Uses MinerU to extract text from images within HTML files and web pages. Features: OCR extra...

Security Scan

The skill's requirements and runtime instructions are consistent with its stated purpose (using the MinerU CLI and a MINERU_TOKEN to OCR HTML), with no obvious mismatches or unexplained privileges.

Purpose & Capability

The skill is an instruction-only wrapper around the mineru-open-api CLI and declares the exact binary and MINERU_TOKEN credential it needs. The declared dependencies (mineru-open-api via npm or Go) and the MINERU_TOKEN credential align with the stated HTML OCR purpose.

Instruction Scope

SKILL.md instructs the agent to run mineru-open-api on local HTML files and to authenticate with MINERU_TOKEN or interactive auth. This stays within the OCR purpose. Note: running the CLI will upload page content/images to MinerU's service (expected for a remote OCR API), so private/sensitive content may be transmitted externally.

Install Mechanism

Install options are npm (mineru-open-api) or go install from a GitHub path. These are expected for a CLI tool and are not downloads from arbitrary URLs. Installing a global npm package or running go install executes third-party code at install time — standard but worth auditing if you don't trust the publisher.

Credentials

The only required environment variable is MINERU_TOKEN (declared as primaryEnv). That is proportionate for an API-backed OCR CLI. No unrelated credentials or extra config paths are requested.

Persistence & Privilege

The skill is user-invocable, not always-on, and does not request special system persistence or cross-skill configuration. It uses normal CLI invocation and environment token-based auth.

Assessment

This skill appears to do what it says: it runs the MinerU CLI and requires a MINERU_TOKEN. Before installing, consider: (1) MINERU_TOKEN is sensitive — only provide it if you trust MinerU/mineru-open-api and the organization. (2) OCRing local HTML/images will likely upload those files to MinerU servers — avoid sending sensitive documents. (3) Global npm packages and go installs execute third-party code at install time; review the mineru-open-api package source or GitHub repo (https://github.com/opendatalab/MinerU) if you need higher assurance. (4) If you decide to proceed, run the CLI in an isolated environment/container and grant the token minimal scope; revoke the token if you suspect misuse.

Use OCR to extract text from HTML files that contain scanned images or image-embedded content using MinerU.

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

# OCR extraction from local HTML file (requires token)
mineru-open-api extract page.html --ocr -o ./out/

# With VLM model for better accuracy
mineru-open-api extract page.html --ocr --model vlm -o ./out/

Authentication

Token required:

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: local .html file
• OCR requires extract with token — not available in flash-extract
• Use --ocr flag to enable OCR on image-embedded content in HTML
• Use --model vlm for complex or mixed-content pages

Notes

• HTML is NOT supported by flash-extract; use extract with token
• If the HTML has normal text content, OCR is not needed — use html-extract instead
• 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