Skip to main content

promptBuilder

codebolt.codeboltutils.PromptBuilder(): InferenceParams
Constructs a smart prompt for the AI agent by combining tools, environment details, system instructions, and task-specific information.

Example​


const codebolt = require('@codebolt/codeboltjs').default;

const { InitialPromptBuilder } = require("@codebolt/codeboltjs/utils");

const promptBuilder = new InitialPromptBuilder(userMessage, codebolt);
const prompt = await promptBuilder
    .addMCPTools()
    .addAgentTools()
    .addEnvironmentDetails()
    .addSystemPrompt('agent.yaml', 'test', 'example.md')
    .addTaskInstruction('task.yaml', 'main_task')
    .buildInferenceParams();

Methods​

The PromptBuilder class provides the following chainable methods:

  • addMCPTools(): Adds MCP (Model Context Protocol) tools to the prompt
  • addAgentTools(): Includes agent-specific tools in the prompt
  • addEnvironmentDetails(): Adds environment context and details
  • addSystemPrompt(yamlFile, section, exampleFile): Loads system instructions from YAML configuration
  • addTaskInstruction(yamlFile, taskName): Adds specific task instructions
  • buildInferenceParams(): Builds and returns the final inference parameters

Usage Notes​

  • All methods return the PromptBuilder instance, allowing for method chaining
  • The constructor requires userMessage and codebolt parameters
  • The final buildInferenceParams() method returns an InferenceParams object ready for LLM execution