Vincent - A secure wallet for your agent

• Outcome: Review
• Latest audit: May 28, 2026 at 6:09 AM UTC
• Version:
  1.0.0

Overview

This is a real crypto-wallet skill whose purpose is clear, but it gives an agent broad ongoing authority to move funds, sign messages, and place bets unless the user locks it down first.

Review before installing. Use a dedicated low-balance wallet, claim it immediately, set spending limits, allowlists, and require-approval policies before funding it, and treat API keys and re-link tokens like financial credentials. Avoid raw signing or arbitrary contract calls unless you personally verify the target, chain, amount, and expected effect.

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 (3)

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This section enables real-value transfers, swaps, arbitrary contract calls, and betting workflows but does not prominently warn that these actions can move or lose real funds irreversibly. In an agent setting, lack of explicit confirmation and risk framing increases the chance of unintended execution, user confusion, or prompt-injected financial actions being carried out.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The skill instructs the agent to store and later retrieve wallet API keys, but does not clearly state that possession of the API key grants transaction authority over the wallet. In practice, this turns the key into a bearer credential for spending and signing operations, so insecure storage or accidental disclosure can directly enable unauthorized transactions.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The re-link flow allows anyone holding a one-time token to obtain a new wallet API key, yet the instructions do not require verification of user authorization or strongly emphasize that the returned API key is highly sensitive. In an agent context, a malicious prompt or untrusted user message containing a token could trick the agent into rebinding wallet access and exposing or using fresh credentials.

VirusTotal

63/63 vendors flagged this skill as clean.