management-deep-dive

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md instructs runtime use of web_search and other public-data tools (e.g., Glassdoor/App Store/merchant forums) to collect “public statements” and “employee/customer/industry feedback,” which are outsider-authored free text that would be ingested into the agent’s LLM context.