agents.planning.llm_compiler_v3.prompts

Prompt templates for LLM Compiler V3 Agent.

Functions

get_executor_prompt(current_task, tool_name, ...)	Generate contextual executor prompt.
get_joiner_prompt(original_query, execution_results, ...)	Generate contextual joiner prompt.
get_planner_prompt(query, available_tools[, scenario])	Generate contextual planner prompt based on scenario.
get_task_fetcher_prompt(completed_tasks, ...[, ...])	Generate contextual task fetcher prompt.

Module Contents

agents.planning.llm_compiler_v3.prompts.get_executor_prompt(current_task, tool_name, resolved_arguments, available_tools)

Generate contextual executor prompt.

Parameters:

• current_task (dict)

• tool_name (str)

• resolved_arguments (dict)

• available_tools (list)

Return type:

str

agents.planning.llm_compiler_v3.prompts.get_joiner_prompt(original_query, execution_results, successful_tasks, failed_tasks)

Generate contextual joiner prompt.

Parameters:

• original_query (str)

• execution_results (list)

• successful_tasks (list)

• failed_tasks (list)

Return type:

str

agents.planning.llm_compiler_v3.prompts.get_planner_prompt(query, available_tools, scenario='default')

Generate contextual planner prompt based on scenario.

Parameters:

• query (str)

• available_tools (list)

• scenario (str)

Return type:

str

agents.planning.llm_compiler_v3.prompts.get_task_fetcher_prompt(completed_tasks, available_tasks, max_parallel, failed_tasks=None)

Generate contextual task fetcher prompt.

Parameters:

• completed_tasks (list)

• available_tasks (list)

• max_parallel (int)

• failed_tasks (list | None)

Return type:

str