Prompt Library Configuration

Overview

LLMling's prompt library allows defining reusable prompts that can be shared across agents. Prompts are defined in the prompts section of your configuration and can be referenced by name.

Basic Structure

prompts:
  # Main system prompts for defining agent behavior
  system_prompts:
    expert_analyst:
      content: |
        You are an expert data analyst.
        Focus on finding patterns and insights.
      type: role

    step_by_step:
      content: |
        Break tasks into sequential steps.
        Explain each step thoroughly.
      type: methodology

# Using prompts in agents
agents:
  analyst:
    system_prompts:           # Direct prompts
      - "You help with analysis."
    library_system_prompts:   # Reference library prompts
      - expert_analyst
      - step_by_step

Prompt Categories

System Prompts

Define core agent behaviors and methodologies:

prompts:
  system_prompts:
    # Role definition
    technical_writer:
      content: |
        You are an expert technical writer.
        Focus on clarity and precision.
        Use proper terminology consistently.
      type: role

    # Methodology definition
    step_by_step:
      content: |
        Follow these steps for each task:
        1. Understand requirements
        2. Plan approach
        3. Execute systematically
        4. Verify results
      type: methodology

    # Tone/style definition
    professional:
      content: |
        Maintain formal, business-appropriate language.
        Be concise but thorough.
      type: tone

Prompt Types

The type field categorizes prompts by their purpose:

Role Types

Define WHO the agent is:

system_prompts:
  expert_dev:
    type: role
    content: |
      You are a senior software developer...

  data_scientist:
    type: role
    content: |
      You specialize in data analysis...

Methodology Types

Define HOW the agent works:

system_prompts:
  analytical:
    type: methodology
    content: |
      Approach problems systematically:
      1. Gather data
      2. Analyze patterns
      3. Form conclusions

  iterative:
    type: methodology
    content: |
      Work in small iterations...

Tone Types

Define communication STYLE:

system_prompts:
  formal:
    type: tone
    content: |
      Use professional language...

  friendly:
    type: tone
    content: |
      Be approachable and helpful...

Format Types

Define output STRUCTURE:

system_prompts:
  markdown:
    type: format
    content: |
      Format responses using Markdown:
      - Use headers for sections
      - Use lists for items
      - Use code blocks for code

Using Library Prompts

Reference prompts in agent configuration:

agents:
  technical_assistant:
    model: gpt-4
    # Direct prompts
    system_prompts:
      - "You are a technical assistant."
      - "Focus on helping with code."

    # Library prompts
    library_system_prompts:
      - technical_writer    # Role
      - step_by_step       # Methodology
      - professional       # Tone
      - markdown          # Format

  data_analyst:
    library_system_prompts:
      - expert_analyst
      - analytical
      - formal

Complete Example

prompts:
  system_prompts:
    # Roles
    technical_expert:
      type: role
      content: |
        You are a technical expert specializing in:
        - Software development

    code_reviewer:
      type: role
      content: |
        You are an experienced code reviewer.
        Focus on:
        - Code quality

    # Methodologies
    systematic:
      type: methodology
      content: |
        Follow this systematic approach:
        1. Understand requirements fully

    # Tones
    professional:
      type: tone
      content: |
        Maintain professional communication:
        - Use formal language
        ...

    # Formats
    structured:
      type: format
      content: |
        Structure responses with:
        1. Clear headings
        2. Bulleted lists
        3. Code examples
        4. Summary points

agents:
  senior_dev:
    model: gpt-4
    description: "Senior developer specialized in code review and optimization"
    system_prompts:
      - "Specialize in Python and TypeScript."
    library_system_prompts:
      - technical_expert
      - code_reviewer
      - systematic
      - professional
      - structured

  teacher:
    model: gpt-4
    description: "Programming teacher and mentor"
    library_system_prompts:
      - technical_expert
      - iterative
      - educational

Organizing Prompts

It's recommended to keep prompt libraries in separate files and use YAML inheritance to include them. This keeps your agent configurations clean and promotes reuse:

# prompts.yml
prompts:
  system_prompts:
    my_prompt:
      content: ...
      type: role

# agents.yml
INHERIT: prompts.yml
agents:
  my_agent:
    library_system_prompts:
      - my_prompt

The integration of this functionality will get improved soon!