​
What is Prompt Tool

Prompt Tool is a component in UR-2.0 used to construct inputs for language models. Each Prompt Tool is defined with the @app.prompt decorator. Its main responsibility is to load template files and generate standardized PromptMessage objects based on inputs such as questions and retrieved passages. These messages can be directly sent to LLM models for generation.

​
How to Implement a Prompt Tool?

Implementing a Prompt Tool usually involves the following three steps:

​
Step 1: Prepare the Prompt Template

Save your prompt template as a file with a .jinja extension, for example:
/images/jinja.svgprompt/qa_rag_boxed.jinja
Please answer the following question based on the given documents.
Think step by step.
Provide your final answer in the format \boxed{YOUR_ANSWER}.

Documents:
{{documents}}

Question: {{question}}

​
Step 2: Implement the Tool in Prompt Server

Use our provided load_prompt_template method to load the template, and implement a tool function (Tool) in the Prompt Server to assemble the prompt:
servers/prompt/src/prompt.py
@app.prompt(output="q_ls,ret_psg,template->prompt_ls")
def qa_rag_boxed(
    q_ls: List[str], ret_psg: List[str | Any], template: str | Path
) -> list[PromptMessage]:
    template: Template = load_prompt_template(template)
    ret = []
    for q, psg in zip(q_ls, ret_psg):
        passage_text = "\n".join(psg)
        p = template.render(question=q, documents=passage_text)
        ret.append(p)
    return ret

​
Step 3: Use the Tool in the Pipeline

  1. Register the Prompt Server and invoke your implemented Tool in the YAML configuration file:
/images/yaml.svgservers/prompt/parameter.yaml
# MCP Server
servers:
  generation: servers/generation
  prompt: servers/prompt

# MCP Client Pipeline
pipeline:
- prompt.qa_rag_boxed
- generation.generate
  1. After building, modify the runtime parameters to specify the path to your template file:
/images/yaml.svgservers/prompt/parameter.yaml
prompt:
  template: prompt/qa_rag_boxed.jinja

​
Bind Different Templates for Multiple Tools

servers/prompt/src/prompt.py
@app.prompt(output="q_ls,ret_psg,template1->prompt_ls")
def qa_rag_boxed(
    q_ls: List[str], ret_psg: List[str | Any], template: str | Path
) -> list[PromptMessage]:
    xxxx

@app.prompt(output="q_ls,ret_psg,template2->prompt_ls")
def qa_rag_boxed_2(
    q_ls: List[str], ret_psg: List[str | Any], template: str | Path
) -> list[PromptMessage]:
    xxxx
Bind template paths to the two tools above by adding the following to servers/prompt/parameter.yaml:
/images/yaml.svgservers/prompt/parameter.yaml
template1: prompt/qa_boxed.jinja
template2: prompt/qa_boxed.jinja
This ensures that different tools use independent templates during execution without interfering with each other.