程序包 io.agentscope.core.hook

类 PostActingEvent

java.lang.Object

io.agentscope.core.hook.HookEvent

io.agentscope.core.hook.ActingEvent

io.agentscope.core.hook.PostActingEvent

public final class PostActingEvent
extends ActingEvent

Event fired after tool execution completes.

Modifiable: Yes - setToolResult(ToolResultBlock)

Context:

• HookEvent.getAgent() - The agent instance
• HookEvent.getMemory() - Agent's memory
• ActingEvent.getToolkit() - The toolkit instance
• ActingEvent.getToolUse() - The original tool call
• getToolResult() - The tool execution result (modifiable)

Note: This is called once per tool execution.

Use Cases:

• Post-process individual tool results
• Filter or sanitize tool output
• Add metadata to results
• Handle tool execution errors
• Transform result format
• Request to stop the agent for human review via stopAgent()

字段概要

从类继承的字段 io.agentscope.core.hook.ActingEvent

toolUse

构造器概要

构造器

构造器	说明
PostActingEvent(Agent agent, Toolkit toolkit, ToolUseBlock toolUse, ToolResultBlock toolResult)	Constructor for PostActingEvent.

方法概要

修饰符和类型	方法	说明
ToolResultBlock	getToolResult()	Get the tool execution result.
Msg	getToolResultMsg()	Get the tool result message.
boolean	isStopRequested()	Check if a stop has been requested.
void	setToolResult(ToolResultBlock toolResult)	Modify the tool execution result.
void	setToolResultMsg(Msg toolResultMsg)	Set the tool result message.
void	stopAgent()	Request to stop the agent after this acting phase.

从类继承的方法 io.agentscope.core.hook.ActingEvent

getToolkit, getToolUse

从类继承的方法 io.agentscope.core.hook.HookEvent

getAgent, getMemory, getTimestamp, getType

从类继承的方法 java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

构造器详细资料

PostActingEvent

public PostActingEvent(Agent agent,
 Toolkit toolkit,
 ToolUseBlock toolUse,
 ToolResultBlock toolResult)

Constructor for PostActingEvent.

参数:

agent - The agent instance (must not be null)

toolkit - The toolkit instance (must not be null)

toolUse - The original tool call (can be null for empty events)

toolResult - The tool execution result (can be null for empty events)

方法详细资料

getToolResult

public ToolResultBlock getToolResult()

Get the tool execution result.

返回:

The tool result block

setToolResult

public void setToolResult(ToolResultBlock toolResult)

Modify the tool execution result.

参数:

toolResult - The new tool result

stopAgent

public void stopAgent()

Request to stop the agent after this acting phase.

When called, the agent will return the current message containing the ToolResultBlock instead of continuing to the next reasoning iteration. The user can then review the tool execution result and resume execution by calling agent.call() with no arguments.

This enables human-in-the-loop scenarios where tool execution results need user review before the agent continues reasoning.

isStopRequested

public boolean isStopRequested()

Check if a stop has been requested.

返回:

true if stopAgent() has been called, false otherwise

getToolResultMsg

public Msg getToolResultMsg()

Get the tool result message.

返回:

The tool result message

setToolResultMsg

public void setToolResultMsg(Msg toolResultMsg)

Set the tool result message.

参数:

toolResultMsg - The tool result message