ClawHub
Back to skill

Security audit

GMGN Skill Track

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed GMGN API helper, with expected credential setup and limited troubleshooting behavior but no evidence of deception, exfiltration, or destructive actions.

Install this only if you intend to use GMGN through gmgn-cli and are comfortable storing a GMGN API key and signing private key locally. Review the npm package trust separately, keep ~/.config/gmgn/.env private, rotate the GMGN key if the machine is shared or compromised, and avoid sharing follow-wallet output publicly because it can reveal your GMGN followed-wallet list.

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 (2)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The skill directs the agent to inspect local IPv6 configuration and query an external IP-check service via shell/network commands. That behavior exceeds the stated wallet-tracking purpose and causes unnecessary host/network reconnaissance and outbound traffic, potentially exposing system metadata or normalizing unrelated environment inspection.

Context-Inappropriate Capability

High
Confidence
97% confidence
Finding
The skill includes first-time setup steps that generate an Ed25519 private key, accept an API key from the user, and persist both secrets to a local config file. This materially broadens the skill from read-only querying into credential generation and secret storage, creating risk of credential leakage, misuse of signing keys, and long-lived compromise if the host or session is exposed.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.