Tool Name: stream_output
Tool Description:
1. Core Function: 
   Real-time output of task execution process/results, generate human-readable natural language content, covering all stages including self-introduction and task execution;
2. Calling Timing (Must be strictly followed):
   - Self-introduction scenario: Call immediately after invoking the self_introduce tool, and only call once;
   - Task execution scenario: Call separately after the completion of each stage: task planning / execution initialization / content generation / result feedback / verification feedback / completion result;
3. Parameter Specifications (Required Fields):
   - content: Fixed stage identifier, only the following values are allowed:
     <Self-introduction> / <Task Planning> / <Execution Process - Initialization> / <Execution Process - Content Generation> / <Execution Process - Result Feedback> / <Verification Feedback> / <Completion Result>;
   - is_final_summary: Only fill in True for the <Completion Result> stage, fill in False for all other stages;
4. Prohibited Scenarios:
   - Repeated calls during the self-introduction stage;
   - Calls when there is no substantive content;
   - Calls when the stage identifier is customized/incorrect.

Input Parameter Examples:
# Self-introduction stage
{"content": "<Self-introduction>", "is_final_summary": false}
# Task planning stage
{"content": "<Task Planning>", "is_final_summary": false}
# Completion result stage
{"content": "<Completion Result>", "is_final_summary": true}