AI Agent Lending - Wallet Credit

• Outcome: Review
• Latest audit: May 29, 2026 at 9:27 PM UTC
• Version:
  2.0.0

Overview

This is a coherent crypto lending skill, but it asks users or agents to grant persistent wallet permissions and perform financial blockchain transactions without enough safety boundaries.

Review carefully before installing. Use only a wallet and funds you can risk, verify the API provider and contract addresses independently, prefer short-lived and low-limit permissions instead of never-expiring permissions, and manually inspect every wallet transaction before signing.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill provides step-by-step instructions for borrowing, approving token spending, and repaying via smart contracts, but it does not clearly warn users that these actions are real on-chain financial transactions that may be irreversible, incur fees, or expose funds to protocol/contract risk. In a lending skill, this omission materially increases the chance that an agent or operator executes transactions without informed consent or adequate human review.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The documentation instructs users to submit wallet addresses to an external API endpoint and references leaderboard lookups, but it does not disclose that wallet addresses and related activity may be logged, correlated, or publicly exposed. While wallet addresses are pseudonymous, linking them to an agent, credit profile, or usage history can create privacy and profiling risks.

VirusTotal

66/66 vendors flagged this skill as clean.