🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
TypeScript analyticsautogenevaluationlangchainlarge-language-modelsllama-indexllmllm-evaluationllm-observabilityllmopsmonitoringobservabilityopen-sourceopenaiplaygroundprompt-engineeringprompt-managementself-hostedycombinator
/**
* MCP Tool: createTextPrompt
*
* Creates a new text prompt version in Langfuse.
*/
import { z } from "zod";
import { defineTool } from "../../../core/define-tool";
import {
CreatePromptSchema,
PromptType,
PromptNameSchema,
COMMIT_MESSAGE_MAX_LENGTH,
PROMPT_NAME_MAX_LENGTH,
} from "@langfuse/shared";
import { createPromptForApi } from "@/src/features/prompts/server/prompt-api-service";
import { buildPromptUrl } from "@langfuse/shared/src/server";
import { runMcpTool } from "../../../core/run-mcp-tool";
import { ParamCreatePromptLabels } from "../validation";
/**
* Base schema for JSON Schema generation (MCP client display)
* Uses simple types that serialize well to JSON Schema
*/
const CreateTextPromptBaseSchema = z.object({
name: z
.string()
.min(1)
.max(PROMPT_NAME_MAX_LENGTH)
.describe("The name of the prompt"),
prompt: z
.string()
.describe("The prompt text content (supports {{variables}})"),
labels: z
.array(z.string())
.optional()
.describe(
"Optional labels to assign, excluding 'production'. New prompt versions receive the 'latest' label automatically.",
),
config: z
.record(z.string(), z.any())
.optional()
.describe(
"Optional JSON config (e.g., {model: 'gpt-4', temperature: 0.7})",
),
tags: z
.array(z.string())
.optional()
.describe("Optional tags for organization (e.g., ['experimental', 'v2'])"),
commitMessage: z
.string()
.optional()
.describe("Optional commit message describing the changes"),
});
/**
* Input schema for runtime validation
* Uses full validation schemas from shared package
*/
const CreateTextPromptInputSchema = z.object({
name: PromptNameSchema,
prompt: z.string(),
labels: ParamCreatePromptLabels,
config: z.record(z.string(), z.any()).optional(),
tags: z.array(z.string()).optional(),
commitMessage: z.string().max(COMMIT_MESSAGE_MAX_LENGTH).optional(),
});
/**
* createTextPrompt tool definition and handler
*/
export const [createTextPromptTool, handleCreateTextPrompt] = defineTool({
name: "createTextPrompt",
description: [
"Create a new text prompt version in Langfuse.",
"",
"Important:",
"- Prompts are immutable - cannot modify existing versions",
"- To update content, create a new version",
"- New prompt versions receive the 'latest' label automatically",
"- Cannot assign the 'production' label during creation",
"- To promote to production, use updatePromptLabels only when the user explicitly requests it",
"- Labels are unique across versions",
"- Use {{variable_name}} syntax for dynamic content",
"",
"Accepts: name, prompt (string), optional labels, config, tags, commitMessage",
].join("\n"),
action: "prompts:CUD",
baseSchema: CreateTextPromptBaseSchema,
inputSchema: CreateTextPromptInputSchema,
handler: async (input, context) => {
return await runMcpTool({
spanName: "mcp.prompts.create_text",
context,
attributes: {
"mcp.prompt_name": input.name,
"mcp.prompt_type": "text",
},
fn: async (span) => {
const createdPrompt = await createPromptForApi({
context,
ctx: context.auth,
input: CreatePromptSchema.parse({
name: input.name,
type: PromptType.Text,
prompt: input.prompt,
labels: input.labels ?? [],
config: input.config ?? {},
tags: input.tags,
commitMessage: input.commitMessage,
}),
});
// Set created version for observability
span.setAttribute("mcp.created_version", createdPrompt.version);
return {
id: createdPrompt.id,
name: createdPrompt.name,
version: createdPrompt.version,
type: createdPrompt.type,
labels: createdPrompt.labels,
tags: createdPrompt.tags,
config: createdPrompt.config,
createdAt: createdPrompt.createdAt,
createdBy: createdPrompt.createdBy,
url: buildPromptUrl({
projectId: context.projectId,
name: createdPrompt.name,
version: createdPrompt.version,
}),
message: `Successfully created text prompt '${createdPrompt.name}' version ${createdPrompt.version}${createdPrompt.labels.length > 0 ? ` with labels: ${createdPrompt.labels.join(", ")}` : ""}`,
};
},
});
},
destructiveHint: true,
});