ClawHub

Arsenal — Sumplus DeFi Execution Layer

Security checks across malware telemetry and agentic risk

Overview

This DeFi skill is coherent, but it gives broad transaction-building guidance for real blockchain activity without tight enough user-confirmation boundaries.

Review before installing. Use only with a limited, revocable Arsenal API key and wallets you are prepared to use for DeFi. Before signing anything, verify the chain, contract, recipient, approval amount, token amounts, slippage, and minimum output in your wallet, and require explicit confirmation for each transaction.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The instruction to invoke Arsenal for anything 'blockchain-related' is overly broad and can cause the agent to activate this skill for read-only, sensitive, or tangential crypto requests without clear user consent. Because this skill can lead to transaction construction and wallet-oriented workflows, broad triggering increases the chance of unintended high-risk actions or unnecessary data sharing to a third-party API.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The manifest description says to use the skill for 'any blockchain operation,' which is an overbroad trigger phrase that can cause unnecessary or unsafe invocation outside the tool's validated scope. In a DeFi execution skill, imprecise activation language is risky because users may be routed into operational flows involving external APIs, addresses, balances, or transaction building when they only wanted information or analysis.

VirusTotal

56/56 vendors flagged this skill as clean.

View on VirusTotal