detect-prompt-injection-mcp-proxy

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). This skill explicitly consumes and pattern-matches tool descriptions from MCP Application Activity (OCSF 1.8 class 6002) "tools/list" response events (normalized by ingest-mcp-proxy-ocsf), which are arbitrary third-party MCP tool metadata that the agent reads and could contain instruction-like/user-generated content capable of influencing tool behavior.