opencr-skill
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a coherent OCR helper, with normal cautions about local file/code access and external OpenOCR model downloads.
This skill looks benign for OCR use. Before installing, confirm you trust the OpenOCR package and any downloaded models, and only provide documents you are comfortable having processed by the agent environment.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent could execute OCR-related Python snippets and access documents supplied for OCR processing.
The skill declares local computer, code execution, and file operation tools. That is expected for OCR workflows that process user-provided images/PDFs, but users should be aware the agent may run local code and read/write task files.
tools: [computer, code_execution, file_operations]
Use the skill only on files you intend to process, and review any proposed commands or file writes before allowing broader local access.
Using the skill may require trusting externally obtained OpenOCR code or model files.
The skill relies on an external OCR library and documents automatic model downloads. This is aligned with OCR functionality, but the artifact does not pin exact package or model versions.
library:\n name: OpenOCR\n url: https://github.com/Topdu/OpenOCR ... auto_download=True, # Auto-download missing models
Install OpenOCR and models from trusted sources, pin versions where possible, and consider specifying local model paths or disabling auto-downloads in sensitive environments.