Unverified50% confidenceFactExact time
科研自动化 Agent 通常基于 LLM 的工具调用与任务规划能力，结合代码执行沙箱、文件系统访问、学术检索及 GitHub/Hugging Face API 等外部工具，形成感知-规划-执行-反馈的闭环
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
Related Claims
UnverifiedAgent-type assistants can receive high-level intent instructions, autonomously plan steps, invoke tools, read and write files, run tests, and complete complex tasks spanning multiple files or even an entire codebase.79% similarUnverifiedIn the Agent architecture, the LLM is responsible for intent understanding and step planning, while actual execution is delegated to external tools such as calculators, search engines, and API endpoints.77% similarUnverifiedAI Agents add a perception-reasoning-action loop on top of LLMs, and are capable of calling external tools such as calendar APIs, email clients, and document editors to complete operations.76% similarUnverifiedAgentic programming tools can understand the overall task objective, autonomously formulate an execution plan, and dynamically adjust strategies based on feedback during execution76% similarUnverifiedSince 2023, projects like Stanford's Generative Agents, Microsoft's AutoGen, and Andrew Ng's Agentic Workflow have demonstrated that LLM-based agents can handle complex cognitive tasks including planning, execution, and reflection76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/500144API
curl https://kongchang.com/api/v1/knowledge/claims/500144MCP
get_claim(id=500144)