OCR - Local (No API Key)

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:20 PM UTC
• Version:
  1.0.0

Overview

This OCR skill appears to do what it says, with the main caveat that “local” processing still includes disclosed first-run language-data downloads.

Install only if you are comfortable with npm dependencies and a first-run download of Tesseract language data. For restricted or offline environments, preload or vendor the language data through a controlled process before use, and avoid OCRing sensitive images unless you are comfortable with the extracted text appearing in the agent session.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The documentation claims the OCR skill is '100% local' while elsewhere stating that Tesseract language data is downloaded at runtime. This is a security-relevant misrepresentation because users may install or run the skill under the assumption that it performs no network access, which can affect trust, sandboxing, compliance, and air-gapped usage decisions.

Intent-Code Divergence

Medium

• Confidence: 95% confidence
• Finding: The feature list says the skill runs purely locally, but the first-run notes disclose downloading language data. This inconsistency can mislead users into granting the skill access in environments where unexpected outbound network activity is prohibited or risky, making it a genuine transparency and supply-chain exposure issue.

VirusTotal

64/64 vendors flagged this skill as clean.