alchemy-agentic-gateway

Pass

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements strong security boundaries for private key management. It explicitly instructs the agent to never use its own file-access tools (Read, Write, Edit) on sensitive files like wallet-key.txt or .env, instead requiring the use of shell pipes and redirects. This prevents the agent from accidentally leaking private keys into its own context or log output.
  • [COMMAND_EXECUTION]: Uses node -e and npx to perform cryptographic operations, such as generating wallets and signing SIWE/SIWS messages. These operations are performed via local shell commands to keep sensitive data isolated from the agent's internal memory.
  • [EXTERNAL_DOWNLOADS]: Fetches necessary libraries and tools (e.g., @alchemy/x402, viem, mppx) from official registries. All remote operations target well-known vendor domains (alchemy.com) or trusted services (Stripe, Circle faucets).
  • [DATA_EXFILTRATION]: No evidence of unauthorized data transfer. Network requests are restricted to the Alchemy gateway infrastructure for the purpose of blockchain API access.
Audit Metadata
Risk Level
SAFE
Analyzed
May 7, 2026, 02:23 PM