Unverified50% confidenceFactExact time
In Agent-as-Tool mode, the expert agent receives only a task string as the user message; its context is deliberately isolated, and the full conversation state of the parent run is not automatically passed in.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
AI Agent Handoff Mechanisms: The Fundamental Difference Between Handoff and Agent-as-Tool
youtubeWhiteboard Bytes10/28/2026
Related Entities
Related Claims
UnverifiedIn the OpenAI Agents SDK, an agent in the tools list is converted into an ordinary tool via the agent.as_tool() method; when called, it starts an independent sub-run and returns only the final output string to the parent agent, and the parent run's conversation state is not inherited.79% similarUnverifiedEach time the Agent calls a tool, it sends only one tool-call request to the server, containing the tool name and the parameters filled in by the AI, with the complete conversation not included.77% similarUnverifiedAgents driven by large language models do not have cross-session memory; at the start of each task, the context contains only the system prompt, the user task, and the files they are asked to read.77% similarUnverifiedIn the context of this article, an agent refers to an AI work unit given clear task boundaries that can independently read and write files; unlike chat-style AI, it is deliberately isolated and receives only a segmented fragment of the truth file and the task for the current chapter, so multiple agents do not interfere with one another.76% similarVerifiedAgent Skill adopts a lazy-loading mechanism whereby the Agent first reads only the metadata (name and description) of each Skill and loads the full instructions only after matching the user's need, so having multiple Skills coexisting does not increase Token consumption76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1107087API
curl https://kongchang.com/api/v1/knowledge/claims/1107087MCP
get_claim(id=1107087)