Contributing

Getting started

git clone https://github.com/pedrotroccoli/1o1-utils.git
cd 1o1-utils
pnpm install

Available scripts

Command	Description
`pnpm test`	Run tests
`pnpm test:coverage`	Run tests with coverage report
`pnpm build`	Compile TypeScript
`pnpm check`	Lint + format check (Biome)
`pnpm check:fix`	Lint + format auto-fix
`pnpm bench`	Run all benchmarks
`pnpm size`	Check bundle sizes

Adding a new utility

1. Create the folder

src/<category>/<util-name>/

Categories: arrays, async, browser, comparisons, formatters, functions, numbers, objects, strings, validators. Use kebab-case for the folder name.

2. Create the files

Every utility has exactly 4 files:

types.ts — Type definitions

interface MyUtilParams {
  // named parameters
}

type MyUtilResult = // return type

type MyUtil = (params: MyUtilParams) => MyUtilResult;

export type { MyUtil, MyUtilParams, MyUtilResult };

index.ts — Implementation

import type { MyUtilParams, MyUtilResult } from "./types.js";

/**
 * Brief description of what this utility does.
 *
 * @keywords common term 1, common term 2, common term 3
 * @see RFC/Standard reference if applicable (https://...)
 */
function myUtil({ param1, param2 }: MyUtilParams): MyUtilResult {
  // validate inputs
  // implementation
}

export { myUtil };

index.spec.ts — Tests (Mocha + Chai)

import { expect } from "chai";
import { describe, it } from "mocha";
import { myUtil } from "./index.js";

describe("myUtil", () => {
  it("should do the expected thing", () => {
    const result = myUtil({ param1: "value" });
    expect(result).to.deep.equal(/* expected */);
  });
});

index.bench.ts — Benchmarks (tinybench)

import { Bench } from "tinybench";
import { getDatasets } from "../../benchmarks/helpers.js";
import { myUtil } from "./index.js";

const bench = new Bench({ name: "myUtil", time: 1000 });

for (const { name, data: getData } of getDatasets()) {
  const data = getData();
  bench
    .add(`1o1-utils (${name})`, () => { myUtil({ /* ... */ }); })
    .add(`lodash (${name})`, () => { /* lodash equivalent */ });
}

export { bench };

3. Register the utility

Add the export to src/index.ts:

export { myUtil } from "./<category>/my-util/index.js";

Add the export entry to package.json:

"./my-util": {
  "import": "./dist/<category>/my-util/index.js",
  "types": "./dist/<category>/my-util/index.d.ts"
}

Add a size-limit entry to .size-limit.json:

{
  "name": "my-util",
  "path": "dist/<category>/my-util/index.js",
  "limit": "1 kB"
}

4. Update `llms.txt` and `llms-full.txt`

Add the new utility to both files at the root of the repository:

llms.txt — add a link entry under the appropriate category section
llms-full.txt — add a full documentation block with signature, parameters, return type, example, and edge cases

These files are served on the docs site for AI tool discoverability.

5. Create a changeset

pnpm changeset

Choose minor for new utilities, patch for bug fixes.

Code conventions

Named object parameters — all functions take a single object, never positional args. Exception: function composition utilities (pipe, once) may use positional/variadic args when it’s the idiomatic pattern.
Type naming — <Name>Params, <Name>Result, <Name> (function type)
No external dependencies — use only TypeScript/JS builtins
Import extensions — always use .js in import paths (required for ESM)
Input validation — validate parameters and throw descriptive errors. Exception: checker functions (isNil, isCircular, isEmpty) return false instead of throwing; safely returns a [error, result] tuple; get returns undefined or a default value.
Named exports only — no default exports
Immutability — all array/object utils must return new instances, never mutate inputs

Naming conventions

Pattern	Convention	Example
Boolean checks	`isX`	`isNil`, `isEmpty`, `isCircular`
Validators	`isValidX`	`isValidUrl`, `isValidEmail`
Conversions	`toX`	`toNumber`
Formatters	`formatX`	`formatSeconds`
Generators	`generateX`	`generateString`
Normalizers	`normalizeX`	`normalizeEmail`
No type suffix	Don’t add the type to the name	`clamp` not `clampNumber`

Discoverability (mandatory)

Every utility must include:

@keywords in JSDoc — common-language terms developers might search for:

/**
 * Clamps a number between a minimum and maximum bound.
 *
 * @keywords limit number, restrict range, min max, bound
 */

@see in JSDoc — RFC or standard reference, if the util aligns with one:

/**
 * @see RFC 5321 — SMTP (https://datatracker.ietf.org/doc/html/rfc5321)
 */

“Also known as” on the docs page — alternative names for the same concept:

Also known as: limit number, restrict range, min/max bound

“Prompt suggestion” on the docs page — a ready-to-copy prompt for AI assistants. Use a fenced text code block (renders a copy button) and always prefix with the canonical references so the LLM can look up the real API:

## Prompt suggestion

```text
I'm using 1o1-utils (npm: https://www.npmjs.com/package/1o1-utils, GitHub: https://github.com/pedrotroccoli/1o1-utils, LLM context: https://pedrotroccoli.github.io/1o1-utils/llms.txt). Show me how to use clamp to restrict a slider value between min and max
```

Style (enforced by Biome)

2-space indentation
Double quotes
Semicolons always
Trailing commas

Run pnpm check:fix before committing.

Testing

Framework: Mocha + Chai
Coverage: c8 with 80% minimum on lines, branches, functions, and statements
Cover happy paths, edge cases, and error cases

Bundle size

Each utility must stay under its size limit (1-2 kB gzipped). The total library limit is 5 kB gzipped. Run pnpm size to verify.