ClawHub
Back to skill

Security audit

test

Security checks across malware telemetry and agentic risk

Overview

This red-team analysis skill is mostly documentation-only, with disclosed local customization and localhost notification behavior that users should understand before enabling.

Install only if you are comfortable with the skill reading its RedTeam customization folder and sending a localhost notification when workflows run. Keep files in that customization directory trusted, and avoid giving the multi-agent workflow secrets or private material you would not want copied into subagent prompts.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (1)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The skill explicitly instructs the agent to issue a background HTTP POST to a localhost service via curl as part of normal execution. Even though the destination is localhost, this expands the skill's capabilities beyond red-team analysis into side-effectful network interaction, which can trigger unintended actions on local services, leak workflow metadata, or be abused in environments where localhost endpoints bridge to privileged host functionality.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.