lifi-mcp-skill
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill connects to a remote MCP server to retrieve tool definitions and execute blockchain queries.
- Evidence: References the endpoint
https://mcp.li.quest/mcpinSKILL.md. - Context: This endpoint belongs to LI.FI, a well-known cross-chain bridge and DEX aggregator.
- [SAFE]: The skill follows security best practices for handling sensitive information.
- Evidence: Uses the
--secret-env LIFI_API_KEYflag inSKILL.mdto avoid hardcoding credentials. - [PROMPT_INJECTION]: The skill processes structured data from an external service, which serves as a potential surface for indirect prompt injection.
- Ingestion points: Data returned from the remote LI.FI MCP endpoint across all provided tool operations.
- Boundary markers: No specific delimiters or boundary instructions are included to isolate external content.
- Capability inventory: Employs the
uxctool to perform network requests for real-time crypto route and balance data. - Sanitization: No explicit sanitization or filtering is applied to the API responses before they are returned to the agent context.
Audit Metadata