Skip to main content

​
Parameters

ParameterTypeDefaultDescription
agentAgentRequiredThe agent instance
taskTaskRequiredThe task being executed

​
Functions

​
get_system_prompt

Public getter to retrieve the constructed system prompt. Returns:
  • str: The final system prompt string

​
manage_system_prompt

Asynchronous context manager for building the system prompt. Parameters:
  • memory_handler (Optional[MemoryManager]): Memory manager for system prompt injection
Returns:
  • AsyncContextManager: Context manager that handles system prompt building
Usage: 
async with system_prompt_manager.manage_system_prompt(memory_handler) as manager:
    system_prompt = manager.get_system_prompt()
    # Use the system prompt
Features:
  • Dynamic Prompt Building: Constructs system prompts from multiple sources
  • Thinking Tool Integration: Handles orchestrated thinking with Thought and AnalysisResult schemas
  • Reasoning Tool Support: Supports both simple and advanced reasoning modes
  • Memory Integration: Integrates with MemoryManager for system prompt injection
  • Agent Context: Includes agent role, goal, instructions, education, and work experience
  • Default Prompt Fallback: Uses default prompt when no agent information is provided
  • Task Context: Handles agent context from task context items
  • Schema Integration: Provides JSON schemas for structured thinking tools
  • Reflective Instructions: Generates comprehensive mission briefing for thinking tools
⌘I