ClawHub
Back to skill

Security audit

PaddleOCR Document Parsing

Security checks across malware telemetry and agentic risk

Overview

This skill is a straightforward PaddleOCR document-parsing helper, with privacy caveats around sending and displaying potentially sensitive documents.

Install only if you trust the PaddleOCR CLI package and are comfortable using a PaddleOCR access token with a hosted document-parsing API. Do not process confidential, regulated, customer, legal, financial, or identity documents unless your data-sharing and retention requirements allow it, and avoid asking the agent to display full extracted text when a summary or redacted output is sufficient.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill encourages users to submit documents via remote URL/API and to save extracted outputs/resources, but it provides no warning about confidentiality, retention, third-party processing, or the sensitivity of uploaded document contents. In a document-parsing skill, inputs commonly include invoices, financial reports, IDs, contracts, or other sensitive records, so omission of privacy/data-handling guidance can lead to unintentional disclosure or policy violations.

Ssd 3

Medium
Confidence
96% confidence
Finding
The instruction to 'always show the full extracted content' promotes disclosure of the entire OCR/parsing output without considering whether the document contains secrets, personal data, or regulated information. In this skill's context, the extracted content is often the user's original sensitive document transformed into machine-readable text, so blanket full disclosure increases the chance of overexposure in chat history, logs, or downstream tooling.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.