cardano-mcp-balances
Pass
Audited by Gen Agent Trust Hub on Mar 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill does not contain any evidence of credential theft, unauthorized data exfiltration, or malicious command execution. It correctly instructs the agent never to ask for sensitive key material.
- [PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it processes data from an external Cardano MCP server. A compromised or malicious server could theoretically return data (such as asset names or metadata) containing hidden instructions.
- Ingestion points: Data retrieved via
get_balances,get_addresses, andget_utxostools as described in SKILL.md. - Boundary markers: Not explicitly defined in the prompt instructions for isolating external tool output.
- Capability inventory: Read-only wallet queries (balances, addresses, UTxOs).
- Sanitization: The skill uses a structured data format for tool outputs, which provides inherent validation against simple text injection.
Audit Metadata