basename-agent

Pass

Audited by Gen Agent Trust Hub on Oct 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill reads data from the system clipboard using navigator.clipboard.readText() to obtain WalletConnect URIs. This represents an attack surface where an attacker-controlled website or application could modify the clipboard content to trick the agent into connecting to a malicious dApp.
  • Ingestion Point: scripts/register-basename.js reads data from the system clipboard via browser automation.
  • Capability Inventory: The scripts include the ability to sign messages and broadcast transactions using a provided PRIVATE_KEY via the ethers library.
  • Boundary Markers: There are no explicit boundary markers or safety instructions provided to the agent regarding the interpretation of the clipboard content beyond a basic protocol check.
  • Sanitization: The script performs a basic check to ensure the URI starts with wc:, but does not validate the destination or content of the WalletConnect session.
  • [COMMAND_EXECUTION]: The skill automates browser interactions using the puppeteer library. While this is essential for its primary purpose of registering Basenames, it involves automated execution of actions within a web browser context driven by external site content.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 4, 2026, 08:43 AM