Unverified50% confidenceFactExact time
Skill invocation is divided into two modes: user-invoked (manually triggered via / commands or $ commands) and model-invoked (automatically determined by the model based on its description).
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
Unverified模型调用技能带有一段description，会进入Agent的上下文中，Agent可自行判断是否调用该技能73% similarUnverified面向客户的Agent只能依赖模型自主触发（model-invoked）的Skills，而无法依赖用户主动调用71% similarUnverified在多模型AI工具中，用户应养成主动确认当前模型的习惯，在重要任务开始前检查模型设置70% similarUnverifiedAgent streaming output presents the model's generation process to the user in real time, including thinking steps, tool invocation processes, and intermediate reasoning states.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/827550API
curl https://kongchang.com/api/v1/knowledge/claims/827550MCP
get_claim(id=827550)