agentic-wallet

Pass

Audited by Gen Agent Trust Hub on Nov 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the npx awal@2.12.1 CLI for wallet management tasks such as authentication (auth login), balance checks (balance), and token transfers (send).
  • [EXTERNAL_DOWNLOADS]: Uses npx to dynamically fetch and run the awal package from the NPM registry. It also provides instructions for installing various official @x402/* and @coinbase/x402 Node.js packages for service monetization.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided inputs like SQL queries, email addresses, and transaction amounts which are interpolated into shell commands.
  • Ingestion points: User-provided SQL for blockchain queries in references/query-onchain.md and transaction details in references/send-usdc.md.
  • Boundary markers: The skill explicitly instructs the agent to use single-quotes and specific JSON structures to delimit user input within shell commands.
  • Capability inventory: Shell command execution via npx and network access to cdp.coinbase.com for blockchain interactions.
  • Sanitization: Each reference file contains a dedicated 'Input Validation' section with specific regex patterns and rules (e.g., rejecting shell metacharacters) to be applied before constructing commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 21, 2026, 08:51 PM