add schema inference feature to trigger config

[nmcgrady]

This PR adds a feature for inferring a response schema based on a mock payload.

Just paste an example payload from your source and select "Infer Schema" and it will automatically add all the values.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
workflow-builder	Ready	Preview	Comment	Nov 27, 2025 7:20am

[vercel]

Suggested change

	field.itemType = arrayStructure.itemType as ValidItemType;
	// Nested arrays case: itemType is "array"
	// Since SchemaField doesn't support itemType: "array",
	// represent nested arrays as itemType: "object" with fields
	if (arrayStructure.itemType === "array") {
	field.itemType = "object";
	field.fields = []; // Empty fields for array of arrays
	} else {
	// Ensure only valid item types are assigned
	field.itemType = arrayStructure.itemType as ValidItemType;
	}

The schema inference can produce invalid SchemaField objects with itemType: "array" when processing nested arrays, but SchemaField only allows itemType to be "string" | "number" | "boolean" | "object".

View Details

Analysis

Schema inference creates invalid SchemaField with itemType: "array" for nested arrays

What fails: inferSchemaFromJSON() in components/workflow/utils/json-parser.ts produces SchemaField objects with itemType: "array", which violates the type contract defined in schema-builder.tsx (line 20) where itemType only allows "string" | "number" | "boolean" | "object".

How to reproduce:

// Call inferSchemaFromJSON with nested arrays
const schema = inferSchemaFromJSON('{"data": [[1, 2], [3, 4]]}');

// Inspect the result
console.log(schema[0].itemType); // Outputs: "array" (INVALID)

What happens: The function assigns itemType: "array" (a string value) through an unsafe as ValidItemType cast on line 101, bypassing TypeScript's type safety. This causes the schema to violate its type contract, and downstream code like template-autocomplete.tsx (line 55) generates invalid type labels like "array[]" instead of valid types.

Expected behavior: For nested arrays, the schema should use valid itemType values. The fix represents nested arrays as itemType: "object" with empty fields, which maintains type safety while properly indicating that array items are complex structures.

Fix implemented: Added a check in createArrayField() to detect when arrayStructure.itemType === "array" and handle it by setting itemType: "object" with an empty fields array, ensuring all SchemaField objects comply with the type contract.

[nmcgrady]

moved to a new PR from my personal account