agent-payment-x402

Pass

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the agentwallet-sdk package from the npm registry. It correctly addresses supply-chain security by recommending explicit version pinning (e.g., @6.0.0) and warning against unpinned npx installs.
  • [COMMAND_EXECUTION]: The skill executes npx agentwallet-sdk as part of its core MCP server functionality. This usage is transparently documented and follows standard practices for model context protocol integrations.
  • [CREDENTIALS_UNSAFE]: Authentication requires a WALLET_PRIVATE_KEY. The skill promotes secure secret management by providing implementation examples that retrieve the key from process environment variables and whitelisting only necessary variables when spawning subprocesses, avoiding hardcoded secrets.
  • [PROMPT_INJECTION]: As the x402 protocol interacts with external server responses, the skill includes robust mitigation strategies for indirect prompt injection. The TypeScript examples implement fail-closed budget enforcement, strict JSON parsing, and validation of numerical data (Number.isFinite) before any financial actions are authorized.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 3, 2026, 10:24 AM