siwa
Pass
Audited by Gen Agent Trust Hub on Mar 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements authentication logic for AI agents using the ERC-8004 standard, facilitating secure identity verification between agents and services.
- [EXTERNAL_DOWNLOADS]: Interacts with official endpoints at api.bankr.bot for transaction submission and siwa.id for documentation. These communications are consistent with the skill's primary purpose and involve a recognized vendor domain.
- [CREDENTIALS_UNSAFE]: Uses environment variables (BANKR_API_KEY, RECEIPT_SECRET) for secret management in its code examples, which is a standard security best practice to prevent hardcoding of sensitive data.
- [PROMPT_INJECTION]: The skill handles agent metadata that is encoded into a URI for blockchain registration. This represents a potential surface for indirect injection, though the implementation follows standard data handling practices.
- Ingestion points: metadata object within references/bankr-signer.md.
- Boundary markers: The data is structured as JSON and encoded into a Base64 URI.
- Capability inventory: Network requests via fetch and blockchain transaction submission.
- Sanitization: Employs Base64 encoding for the metadata payload before inclusion in transactions.
Audit Metadata