OCR Locally

PassAudited by VirusTotal on May 4, 2026.

Overview

Type: OpenClaw Skill Name: ocr-locally Version: 1.0.0 The skill provides local OCR (Optical Character Recognition) capabilities for macOS using native Apple frameworks (Vision, PDFKit, and AppKit). The Swift scripts (ocr_vision_pro.swift, pdf_ocr.swift) are well-documented and perform text extraction entirely on-device without any network requests, data exfiltration, or suspicious shell execution. The instructions in SKILL.md are strictly aligned with the stated purpose of processing images and PDFs provided by the user.

Findings (0)

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.

What this means

Users may not realize from the registry metadata that installing the skill includes local code that will be run with Swift on macOS.

Why it was flagged

The package metadata under-declares the runnable Swift scripts and runtime dependency, even though the manifest includes scripts and the SKILL.md instructs users to run them. This is a provenance/packaging clarity issue, not evidence of hidden malicious behavior.

Skill content
Install specifications: No install spec — this is an instruction-only skill. Code file presence: No code files present — this is an instruction-only skill. ... File manifest: ... scripts/ocr_vision_pro.swift ... scripts/pdf_ocr.swift
Recommendation

Treat this as a local-code skill: review the included Swift scripts, ensure Swift/macOS requirements are acceptable, and prefer metadata that declares the macOS and swift runtime requirements.

What this means

The skill will run code on the user's Mac to process files the user provides.

Why it was flagged

The skill's normal operation executes local Swift scripts. This is purpose-aligned for local OCR and the provided code uses native macOS OCR/PDF APIs, but it is still local code execution.

Skill content
swift scripts/ocr_vision_pro.swift <image_path> ... swift scripts/pdf_ocr.swift <pdf_path>
Recommendation

Run it only on intended files, keep prompts user-directed, and inspect the scripts before first use if installing from an unknown source.

What this means

OCR results may be saved to local files, and existing files at the chosen output path may be replaced.

Why it was flagged

The OCR script can write extracted text and confidence details to user-specified paths. This is expected for an OCR tool, but output paths could overwrite files if chosen carelessly.

Skill content
try text.write(toFile: absolutePath, atomically: true, encoding: .utf8)
Recommendation

Use explicit safe output paths, avoid protected or important filenames, and quote/validate user-provided paths when constructing commands.