ClawHub
Back to skill

Security audit

Ethereum Read Only

Security checks across malware telemetry and agentic risk

Overview

This skill is a read-only Ethereum command guide with some privacy and shell-safety caveats, but no evidence of hidden wallet access, fund movement, destruction, or exfiltration.

Install only if you are comfortable running Foundry's official installer and sending read-only blockchain queries to your chosen RPC/API providers. Use a dedicated low-privilege RPC key or a trusted/self-hosted node for sensitive investigations, avoid entering private keys or seed phrases, and treat the external wallet/transaction guide as outside this skill's reviewed read-only scope.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill is presented as a read-only Ethereum state inspection guide, but the ABI decoding example silently queries an external web service (4byte.directory) with transaction-derived selectors. This expands the trust boundary beyond the stated RPC-only blockchain reads, creating privacy, integrity, and availability risks because user analysis activity and queried selectors may be disclosed to a third party and results may be spoofed or unavailable.

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The event log decoding workflow also calls 4byte.directory even though the skill is framed as on-chain read-only access via Foundry cast. This can leak which contracts, logs, or investigations the user is analyzing and introduces an undeclared dependency whose responses could mislead analysis if incorrect or manipulated.

Description-Behavior Mismatch

Low
Confidence
81% confidence
Finding
The closing note directs users of a read-only analysis skill to an unrelated external site for wallet management and transaction guidance, which is outside the advertised scope. Even without direct code execution, this creates a trust pivot that could steer users toward higher-risk actions or unreviewed content, especially in a blockchain context where wallet operations are security-sensitive.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.