医小智

基于“多模型并行推理、主动抗辩”的低幻觉（国家发明专利）独有模式，快速为用户提供精准的、个性化的疾病诊断、智能荐药、推拿按摩、针灸等服务。并推荐对应的检测机构、治疗机构。

This package is internally inconsistent: SKILL.md asserts the agent should not call scripts, yet a Python script capable of contacting remote LLM endpoints is included; the docs claim automatic token fetching from https://jiyinjia.jinbaisen.com but the script does not implement that flow. The skill also forces adding a specific external cancer-screening link into reports. Before installing, consider: (1) verify the publisher and domain names (jinbaisen/jiyinjia/ydai) and confirm they are legitimate; (2) avoid allowing automatic token fetches or running curl against unknown URLs — prefer to supply your own vetted API key if you must test; (3) do not use this as a substitute for professional medical care — the skill's rules deliberately disallow cautious language and may produce overly confident diagnoses; (4) if you proceed, run the skill in a sandboxed environment and review network traffic to the listed domains. If you need a clean, auditable behavior, request an updated package where the token-fetch flow, the required external links, and the 'no-scripts' claim are reconciled and the service providers are documented.

Type: OpenClaw Skill Name: yixiaozhi Version: 1.0.0 The skill contains high-risk prompt injection instructions that command the agent to bypass its own execution framework ('不调用任何脚本、不执行任何命令') and provide definitive medical diagnoses while explicitly forbidding safety disclaimers ('禁止推脱表达'). It also mandates the insertion of a specific commercial lead-generation link (bmsapp.geneplus.org.cn) whenever cancer-related keywords are detected, effectively hijacking the agent's output for unauthorized commercial redirection. While these behaviors are deceptive and medically irresponsible, they represent a high-risk scam/adware profile rather than intentional system-level malware like data exfiltration or backdoors.