cmc-mcp

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill provides configuration instructions for the CoinMarketCap MCP server at 'https://mcp.coinmarketcap.com/mcp'. This represents a network operation to a well-known service required for the skill's primary functionality. No sensitive file access or unauthorized data exfiltration patterns were detected.
  • [PROMPT_INJECTION]: The skill is designed to ingest external data from cryptocurrency news and project descriptions, which creates a surface for indirect prompt injection. Evidence Chain: 1. Ingestion points: External content is retrieved via the 'mcp__cmc-mcp__get_crypto_latest_news' and 'mcp__cmc-mcp__get_crypto_info' tools in 'SKILL.md'. 2. Boundary markers: The skill instructions do not define delimiters or ignore-embedded-instruction markers for this external data. 3. Capability inventory: The skill's capabilities are restricted to information retrieval tools; it does not have access to command execution, file system modification, or non-whitelisted network operations. 4. Sanitization: There is no mention of sanitization, filtering, or validation of the retrieved external text.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 08:46 AM
Security Audit — agent-trust-hub — cmc-mcp