PromptTuner MCP

// Zod input schemas for PromptTuner MCP import { z } from 'zod'; import { MAX_PROMPT_LENGTH } from '../config/constants.js'; import { OPTIMIZATION_TECHNIQUES, TARGET_FORMATS } from '../config/types.js'; const promptSchema = z .string() .min(1, 'Prompt cannot be empty') .max( MAX_PROMPT_LENGTH, `Prompt exceeds maximum length of ${MAX_PROMPT_LENGTH} characters` ) .describe('Prompt text to improve (plain text, Markdown, or XML)'); const techniqueSchema = z .enum(OPTIMIZATION_TECHNIQUES) .describe( 'basic | chainOfThought | fewShot | roleBased | structured | comprehensive' ); const targetFormatSchema = z .enum(TARGET_FORMATS) .describe('auto | claude | gpt | json'); export const RefinePromptInputSchema = { prompt: promptSchema, technique: techniqueSchema .optional() .default('basic') .describe( 'basic | chainOfThought | fewShot | roleBased | structured | comprehensive' ), targetFormat: targetFormatSchema .optional() .default('auto') .describe('auto | claude | gpt | json'), }; export const AnalyzePromptInputSchema = { prompt: promptSchema, }; export const OptimizePromptInputSchema = { prompt: promptSchema, techniques: z .array(techniqueSchema) .min(1, 'At least one technique required') .max(6, 'Maximum 6 techniques allowed') .default(['basic']) .describe( 'Array of: basic, chainOfThought, fewShot, roleBased, structured, comprehensive' ), targetFormat: targetFormatSchema .optional() .default('auto') .describe('auto | claude | gpt | json'), }; export const DetectFormatInputSchema = { prompt: promptSchema, }; export const ComparePromptsInputSchema = { promptA: promptSchema.describe('First prompt to compare'), promptB: promptSchema.describe('Second prompt to compare'), labelA: z .string() .optional() .default('Prompt A') .describe('Label for first prompt'), labelB: z .string() .optional() .default('Prompt B') .describe('Label for second prompt'), }; export const ValidatePromptInputSchema = { prompt: promptSchema.describe('Prompt to validate'), targetModel: z .enum(['claude', 'gpt', 'gemini', 'generic']) .optional() .default('generic') .describe('Target AI model for token limit validation'), checkInjection: z .boolean() .optional() .default(true) .describe('Check for prompt injection patterns'), };