The Prompt Injection vulnerability present in Coinbase AgentKit has been addressed, but the actual impact has been significantly underestimated.

robot
Abstract generation in progress

According to CriptoNoticias, an independent security researcher disclosed that Coinbase AgentKit has a prompt injection vulnerability, allowing attackers to induce the AI agent to execute unauthorized token transfers without manual confirmation. This vulnerability has been verified through actual transactions on the Base Sepolia testnet. Additionally, the researcher pointed out that the flaw also exposes an infinite approval process for ERC-20 tokens and access permissions to remote servers within the same agent execution context, extending the risk beyond wallet depletion, though the report did not specify which infrastructure components might be affected. The vulnerability was submitted to Coinbase’s bug bounty program in February and officially verified, ultimately classified as medium risk with a $2,000 bounty paid. However, the researcher emphasized that the actual impact of the vulnerability is much greater than the official rating suggests.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin