Business License Ocr

• Outcome: Review
• Latest audit: May 29, 2026 at 9:33 PM UTC
• Version:
  1.0.4

Overview

This appears to be a legitimate cloud OCR skill, but its instructions are broader than its business-license purpose and could send sensitive documents to an external OCR API.

Install only if you intend to send selected documents to Scnet for OCR. Before use, confirm the file path, document type, and endpoint, and avoid routing IDs, invoices, or other sensitive documents through this skill unless that broader processing is intentional and acceptable under your privacy requirements.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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

Findings (4)

Description-Behavior Mismatch

Medium

• Confidence: 86% confidence
• Finding: The changelog states the skill supports 10 OCR document types, while the manifest describes it as specifically for Chinese mainland business licenses. This capability mismatch can mislead users and reviewers about the actual data the skill can process, expanding the effective attack surface to additional sensitive document classes such as IDs, bank cards, and invoices.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The manifest says the skill is for business-license OCR, but the documentation advertises broader OCR for IDs, invoices, and generic text. This scope expansion can cause the agent to invoke the skill on more sensitive documents than users expect, increasing privacy and compliance risk when those files are sent to a third-party service.

Intent-Code Divergence

Medium

• Confidence: 90% confidence
• Finding: The examples and tags contradict the declared single OCR type by suggesting use for identity documents, invoices, and generic text. In practice, this can broaden automatic triggering and lead to unintended handling of personal or regulated data outside the skill's stated purpose.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The AI trigger guidance is overly broad, encouraging automatic invocation for many OCR-like requests that do not match the documented capability. Over-broad triggering is dangerous because it can silently route unrelated sensitive files to an external OCR API without sufficiently specific user intent.

VirusTotal

65/65 vendors flagged this skill as clean.