Build and revise procedural 3D assets with your coding agent. Local MCP server, CLI and TypeScript engine with rendered review, editable source and GLB export.
JavaScript 3d3d-assets3d-modelingagent-skillsagent-toolsai-agentsasset-pipelinecligame-assetsglbgltfmcpmcp-servermodel-context-protocolprocedural-generationprocedural-modelingtext-to-3dthreejstypescript
/**
* `docs/tools.md` is generated, and `docs/` is in the package's `files` list, so a
* stale reference is published to every install. The generator has had a `--check`
* mode from the start; it appeared in no workflow and no test, which is the same
* shape as the render service's 37 tests running nowhere (13.7).
*
* What slipped through it: 13.2's in-range `zod` 4.4.3 -> 4.6.2 bump changed
* `z.toJSONSchema` to emit `items: false`, `minItems` and `maxItems` for
* fixed-length tuples, so every tuple input in the published reference understated
* its own schema. An offline gate that exists but runs nowhere is worth exactly as
* much as no gate.
*/
import { expect, test } from 'bun:test';
import { readFile } from 'node:fs/promises';
import { toolReferenceMarkdown, toolReferencePath } from './generate-tool-reference.ts';
import { operationReferenceMarkdown, operationReferencePath } from './generate-tool-reference.ts';
test('docs/tools.md matches the registry it is generated from', async () => {
const published = await readFile(toolReferencePath, 'utf8');
// Compared whole rather than by section: the drift that shipped was four added
// lines inside one nested schema, which any summary comparison would have missed.
expect(
published === toolReferenceMarkdown(),
'docs/tools.md is stale — run `bun run docs:tools`',
).toBe(true);
});
test('named operation reference matches the registry without changing default tools', async () => {
expect(await readFile(operationReferencePath, 'utf8')).toBe(operationReferenceMarkdown());
expect(operationReferenceMarkdown().match(/^## kiln_/gm)).toHaveLength(21);
});