ClawHub

Image To Data

Security checks across malware telemetry and agentic risk

Overview

This skill is purpose-aligned for analyzing construction images, but users should treat uploaded images and extracted text as potentially sensitive because external vision APIs may be used.

Install only if your organization allows construction photos, scans, drawings, and extracted text to be processed by the configured AI provider. Use scoped provider API keys, avoid exposing unrelated credentials in the environment, and redact sensitive site, client, personnel, or project details when possible.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill explicitly processes site photos, scanned documents, drawings, and OCR output, all of which can contain sensitive information such as worker identities, project details, schedules, or proprietary plans. The documentation provides usage patterns for loading and exporting this data but does not warn about sensitive-data handling, retention, redaction, or access controls, which increases the chance of inadvertent exposure in real deployments.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The description is very broad ('Extract data from construction images using AI Vision. Analyze site photos, scanned documents, drawings.') and does not define clear scope limits, disallowed content, or activation conditions. In combination with declared filesystem and network permissions, this ambiguity can cause the skill to be invoked on sensitive images or documents more often than intended, increasing the risk of over-collection, unintended exfiltration, or unsafe downstream handling.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal