add-guard-protection

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and language references explicitly instruct integrating detectPromptInjection to scan tool call results (e.g., the "tools.search_web" example and the "Scanning tool results for injection — detectPromptInjection" entry), meaning the agent will ingest untrusted third-party/tool/web content and let those results influence guard decisions that can change behavior.