skills/jeremylongshore/tons-of-skills-marketplace/alchemy-reference-architecture/Gen Agent Trust Hub
alchemy-reference-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or data exfiltration techniques were identified. The skill content consists of architectural guidelines and project structure definitions aimed at secure development.- [SAFE]: The guidelines explicitly promote a security-first approach, directing the agent to implement API key proxies and managed secrets to prevent credential exposure to frontend components.- [INDIRECT_PROMPT_INJECTION]: The skill describes an architecture designed to process untrusted external data.
- Ingestion points: Wallet addresses and webhook event payloads enter the system at the API Layer (Next.js/Express).
- Boundary markers: The skill specifies using signed challenges for ownership proof and HMAC signature verification for webhooks.
- Capability inventory: The skill uses Read, Write, and Edit tools to define the project structure and infrastructure logic.
- Sanitization: The instructions mandate early rejection of malformed inputs and the use of dedicated validation logic for blockchain-specific data (addresses, blocks).
Audit Metadata