ClawHub

Clawdwallet

Security checks across malware telemetry and agentic risk

Overview

This is a real wallet-control skill, but it gives an agent broad authority over mnemonics and transaction signing with limited safeguards.

Install only if you are comfortable reviewing and building the referenced wallet code yourself. Use a new low-value wallet, keep the WebSocket bound to localhost on a trusted machine, require manual human confirmation for every signature or transaction, and never provide a primary wallet mnemonic.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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

High
Confidence
91% confidence
Finding
The skill instructs an agent to initialize a wallet from a mnemonic and approve signing requests, but it does not prominently warn that signatures and transactions can irreversibly transfer assets, grant token approvals, or expose users to unauthorized wallet actions. In an agent-controlled wallet context, missing explicit consent and risk language materially increases the chance that an operator will approve dangerous requests without understanding the consequences.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill describes a WebSocket-controlled browser wallet using a local endpoint, but it does not clearly warn that any local process, malicious site-assisted local malware, or misconfiguration could abuse that control channel to drive wallet actions. Because this extension is explicitly designed for agent control, the exposed control surface is unusually sensitive and should be treated like remote wallet administration.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal