赤脚医生智能助手

This skill appears internally consistent with its stated purpose (a local, reference-based medical assistant). Before installing or enabling it, consider: - Medical risk: the references include concrete, actionable procedures (acupuncture point locations and needling depths, emergency maneuvers). Even with disclaimers, non-professional users could follow instructions and be harmed. Only expose to users who understand the limitations. - Liability & regulation: providing medical guidance may have legal/regulatory implications in your jurisdiction; ensure usage policies and disclaimers meet local requirements. - Autonomy: the skill can be invoked autonomously by the agent by default; if you are uncomfortable with an agent acting on medical queries without human review, consider disabling autonomous invocation for this skill. - Testing: review outputs in a safe environment and have clinicians validate recommendations before relying on them. - Audit: because the skill contains no network code, it does not exfiltrate data by itself — nonetheless, monitor logs and vet any integration you add (e.g., telemetry or connectors) to avoid adding remote endpoints later.