prospect

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

• Third-party content exposure detected (high risk: 0.85). The skill calls multiple runtime MCP tools (company search, bulk enrich, people search, bulk match) that fetch and return third-party Apollo web/registry data (e.g., company/person records and emails) into the agent’s LLM context, which is outsider-authored free text/fields not authored by the operating user.