OSSS.ai.workflows.prompt_loader¶
OSSS.ai.workflows.prompt_loader
¶
Prompt loading utilities for workflow execution.
This module provides functions to load prompts from agent prompts.py files and apply custom prompt configurations from workflow definitions.
load_agent_prompts(agent_type)
¶
get_system_prompt(agent_type, custom_prompt=None)
¶
Get the system prompt for an agent, with optional custom override.
Parameters¶
agent_type : str The type of agent (e.g., 'refiner', 'critic', 'historian', 'synthesis') custom_prompt : Optional[str] Custom prompt to override the default system prompt
Returns¶
Optional[str] The system prompt to use, or None if no prompt is available
get_template_prompt(agent_type, template_name, **kwargs)
¶
Get and format a template prompt for an agent.
Parameters¶
agent_type : str The type of agent (e.g., 'refiner', 'critic', 'historian', 'synthesis') template_name : str The name of the template prompt to retrieve **kwargs Variables to substitute in the template
Returns¶
Optional[str] The formatted template prompt, or None if not found
apply_rich_configuration_to_prompts(agent_type, config)
¶
Apply rich configuration options to generate customized system prompts.
Maps configuration options like refinement_level, search_depth, etc. to prompt modifications that control agent behavior.
Parameters¶
agent_type : str The type of agent (refiner, historian, critic, synthesis) config : Dict[str, Any] Rich configuration dictionary with agent-specific options
Returns¶
str Customized system prompt based on configuration
apply_prompt_configuration(agent_type, config)
¶
Apply prompt configuration from workflow node configuration.
Supports both simple custom prompts and rich configuration options.
Parameters¶
agent_type : str The type of agent config : Dict[str, Any] Configuration dictionary that may contain prompt overrides
Returns¶
Dict[str, str] Dictionary containing the configured prompts