insurance-policy-faq-explainer

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

• Third-party content exposure detected (high risk: 0.75). 该技能在运行时会接收用户提供的 insurerPageLink（公开可访问 URL）并将其对应页面/条款内容作为可读文本用于上下文拼装；该页面属于“公共网页内容/运行时抓取”的外部来源，存在间接提示注入风险。

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The scripts/run.py runner makes runtime API calls to the external execution endpoint at https://ai-skills.ai (DEFAULT_BASE_URL) which is required for executing the skill and thus fetches remote responses that directly control the agent's prompts/behavior.