NYC

authenticate-wallet

Fail

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • REMOTE_CODE_EXECUTION (HIGH): The skill relies on npx awal@latest for its core functionality. npx fetches the specified package from the npm registry and executes it immediately. Since the 'awal' package is not from a trusted organization, this mechanism can be used to execute arbitrary malicious code on the system. Using the @latest tag further increases risk as the code can change at any time without a lockfile.\n- PROMPT_INJECTION (HIGH): The skill exhibits a high surface for Indirect Prompt Injection. \n
  • Ingestion points: The skill explicitly instructs the agent to read the user's email to retrieve OTP codes (SKILL.md). \n
  • Boundary markers: No markers or safety instructions are provided to help the agent distinguish between the OTP code and potentially malicious instructions embedded in the email. \n
  • Capability inventory: The agent is granted the ability to run multiple CLI commands through Bash, including wallet authentication and balance checks. \n
  • Sanitization: There is no mention of sanitizing or validating the email content before processing. An attacker could send an email that tricks the agent into performing unauthorized wallet operations.\n- EXTERNAL_DOWNLOADS (MEDIUM): The skill downloads and runs an unpinned, third-party package from an external registry (npm).
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 15, 2026, 08:31 PM