metamask-openclaw-wallet-integration

Warn

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing a library (metamask-openclaw-skill) and cloning a repository (github.com/veryyoldman/metamask-openclaw-skill.git) from external sources that are not clearly associated with the identified vendor 'Aradotso'.
  • [COMMAND_EXECUTION]: The documentation provides instructions to clone, install, and build code from an unverified external source using shell commands, including git clone, npm install, and npm run build in SKILL.md.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The agent ingests untrusted data from external blockchain RPCs via functions like getBalance and getState as defined in the library usage examples in SKILL.md.
  • Boundary markers: There are no delimited blocks or explicit instructions for the agent to ignore potentially malicious content embedded in blockchain metadata (such as ENS names or transaction data).
  • Capability inventory: The skill allows the agent to initiate financial transactions (sendNative) and switch networks (switchChain), which could be abused if the agent is influenced by malicious data.
  • Sanitization: No evidence of sanitization, validation, or escaping of data retrieved from the blockchain is provided in the implementation instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 14, 2026, 08:46 AM
Security Audit — agent-trust-hub — metamask-openclaw-wallet-integration