ClawHub

NexSolve AI 需求广场

Security checks across malware telemetry and agentic risk

Overview

This skill appears to publish user-provided reports, and optional contact details, to a public GitHub issue without a separate confirmation step.

Install only if you are comfortable with report text and contact details being sent to GitHub and potentially becoming public. Users should be warned before submission, avoid secrets or private business details, and confirm the destination repository and visibility before allowing the skill to create an issue.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill sends user-supplied problem descriptions and optional contact information directly to GitHub as a public issue, which can expose sensitive business details or personal data to a third-party service and the public repository. Although the issue body contains submission terms, the code does not present an explicit runtime consent flow or validation to prevent accidental disclosure of confidential information.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill sends user-supplied problem descriptions and optional contact information to a public GitHub repository, which can expose sensitive personal or business data to third parties. Although the issue body contains text asserting that the submission will be public, that notice is embedded in the payload sent to GitHub rather than enforced as a clear user-facing consent gate before transmission, so users may disclose data without meaningful informed consent.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal