agentleFS
crustacean-devcrustacean-dev/stack-guardrails/skills/nodejs/SKILL.md

nodejs

Skill0 starsChanged 8 months agoMITView on GitHub

ALWAYS consult this skill before writing, editing, reviewing, or refactoring any Node.js code. It prevents the most common LLM-generated Node.js mistakes: bare 'fs'/'path' imports without the node: prefix, sync APIs (readFileSync) in async contexts, exec() shell injection, JSON.parse(JSON.stringify()) instead of structuredClone(), Math.random() instead of crypto.randomUUID(), missing graceful shutdown, and callback-style APIs when promise versions exist. Trigger on: any server-side JavaScript/TypeScript, CLI tools, scripts, Express/Fastify/Koa apps, anything importing fs/path/crypto/child_process/http/stream/worker_threads, package.json edits, require()-to-ESM refactors, or any mention of Node.js, npm, or server-side JS. Also trigger when you see code with require(), readFileSync, exec(), JSON.parse(JSON.stringify()), or Math.random() in a Node.js context — the skill tells you what to replace them with. Do NOT trigger for browser-only JS, React/Vue/Svelte components, Deno, or Bun-specific code.

7 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use grpc, gcp and 4 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Sorcha-PlatformSkill

    grpc

    Build or review gRPC services and clients in .NET. USE FOR: ASP.NET Core gRPC, protobuf contracts, unary or streaming RPC, gRPC client factory, interceptors, deadlines, cancellation, channel reuse, backend service integration. DO NOT USE FOR: broad browser-facing APIs without gRPC-Web tradeoff review, SignalR realtime hubs, plain REST APIs. INVOKES: dotnet build/test and focused service or client smoke checks when code changes.

    Also in 3 projects that use this

  • ffsshhttiikkSkill

    gcp

    Google Cloud Platform services and best practices

    Also in 3 projects that use this

  • VercelSkill

    oauth

    Configure OAuth providers (Google, Apple, Microsoft, Facebook, GitHub, etc.) to work with portless local dev URLs. Use when setting up OAuth redirect URIs, fixing "redirect_uri_mismatch" or "invalid redirect" errors, configuring sign-in providers for local development, or when a provider rejects .localhost subdomains. Triggers include "OAuth not working with portless", "redirect URI mismatch", "Google/Apple/Microsoft sign-in fails locally", "configure OAuth for local dev", or any task involving OAuth callback URLs with portless domains.

    Also in 3 projects that use this

  • JetBrainsSkill

    vue

    Vue 3 Composition API, script setup macros, reactivity system, and built-in components. Use when writing Vue SFCs, defineProps/defineEmits/defineModel, watchers, or using Transition/Teleport/Suspense/KeepAlive.

    Also in 4 projects that use this

  • MicrosoftSkill

    github-actions

    Use when creating, editing, reviewing, or troubleshooting GitHub Actions workflows in Scope.

    Also in 4 projects that use this

  • VercelSkill

    react

    React renderer for json-render that turns JSON specs into React components. Use when working with @json-render/react, building React UIs from JSON, creating component catalogs, or rendering AI-generated specs.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    No .then() chains: when async/await is available.

    line 29
  2. Very specific:

    Promise.allSettled() over Promise.all(): when partial failure is acceptable — Promise.all() rejects on the first failure and you lose …

    line 30

“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Node.js Guardrails

These guardrails apply to all Node.js code you write, review, or modify. They exist because LLM-generated Node.js tends to repeat the same mistakes: missing node: prefixes, callback-style APIs when promise versions exist, unsafe exec() calls, sync fs operations in async contexts, and JSON.parse(JSON.stringify()) for deep copies. Following these rules produces code that's modern, secure, and idiomatic for Node.js 20+.


Imports

The node: prefix makes it unambiguous that you're importing a built-in module, not a npm package that happens to share the name. Node.js has had this prefix since v16 and it should be used everywhere.

  • Always use the node: prefix for built-in modules — import fs from 'node:fs/promises', not import fs from 'fs'. This applies to every built-in: node:path, node:url, node:crypto, node:os, node:stream, node:child_process, node:util, node:events, etc.
  • Use node:fs/promises — not node:fs with callbacks. The promise-based API is cleaner and composes with async/await.
  • Use node:path for all path operations — never concatenate strings with /. Paths are OS-dependent and path.join() / path.resolve() handle this correctly.
  • Use node:url and the URL class for URL manipulation — not string concatenation or regex.
  • Use node:crypto for randomness — never Math.random() for anything security-related (tokens, IDs, secrets). crypto.randomUUID() and crypto.randomBytes() exist for this reason.

Async

Node.js is built around non-blocking I/O. Callbacks were the original API style, but modern Node.js has promise-based alternatives for everything. Using async/await makes code readable, debuggable, and composable.

  • async/await everywhere — never callbacks unless forced by a legacy API that has no promise version.
  • No .then() chains when async/await is available. .then() chains are harder to read, harder to debug (stack traces), and harder to compose with try/catch.
  • Promise.allSettled() over Promise.all() when partial failure is acceptable — Promise.all() rejects on the first failure and you lose the results of the others.
  • AbortController + AbortSignal for cancellation — not custom boolean flags like let cancelled = false. AbortSignal is the standard cancellation mechanism and integrates with fetch, streams, child processes, and timers.
  • Use node:timers/promises — import { setTimeout } from 'node:timers/promises' gives you an awaitable timer directly. Don't wrap setTimeout in new Promise().
// wrong
await new Promise(resolve => setTimeout(resolve, 1000));

// correct
import { setTimeout } from 'node:timers/promises';
await setTimeout(1000);

Process

  • No process.exit() except in CLI entry points — in library code and most application code, throw an error or return an error value instead. process.exit() skips cleanup, kills pending I/O, and makes code untestable.
  • Use process.exitCode = 1 over process.exit(1) when possible — setting the exit code lets Node.js finish pending operations before exiting naturally.
  • Handle SIGINT and SIGTERM in long-running processes (servers, workers, daemons) for graceful shutdown — close connections, flush buffers, release resources.
  • structuredClone() for deep copy — not JSON.parse(JSON.stringify()). The JSON trick silently drops undefined, functions, Date objects, Map, Set, RegExp, and circular references. structuredClone() handles all of these correctly.
  • Use node:worker_threads for CPU-heavy work — not child_process for running JS code. Worker threads share memory (via SharedArrayBuffer) and avoid the overhead of spawning a new process.

File System

Sync fs operations block the event loop. In a server or any async context, this means every other request stalls while you wait for disk I/O. The only acceptable place for sync fs is CLI startup where nothing else is running yet.

  • Always node:fs/promises — never sync fs operations (readFileSync, writeFileSync, etc.) except during CLI startup/initialization.
  • Specify encoding explicitly — readFile(path, { encoding: 'utf-8' }), not readFile(path) which returns a Buffer.
  • mkdir with { recursive: true } — don't check existence with existsSync first and then create. The recursive option is idempotent and avoids race conditions.
  • Use node:os tmpdir() for temp files — never hardcode /tmp. macOS uses /private/var/folders/..., Windows uses %TEMP%, and containers may mount tmpfs elsewhere.
import { mkdir, readFile, writeFile } from 'node:fs/promises';
import { tmpdir } from 'node:os';
import { join } from 'node:path';

await mkdir(join(tmpdir(), 'my-app'), { recursive: true });
const content = await readFile(configPath, { encoding: 'utf-8' });

Streams

Streams are powerful but notoriously error-prone when piped manually. A leaked error handler means an unhandled rejection that crashes your process. pipeline() handles backpressure, error propagation, and cleanup for you.

  • Use pipeline() from node:stream/promises — never .pipe() with manual error handling. pipeline() destroys all streams on error and returns a promise.
  • Prefer node:stream/consumers (.text(), .json(), .buffer()) for consuming readable streams — these are built-in and handle encoding correctly.
  • Use Readable.from() for creating streams from iterables — don't push data manually into a PassThrough stream.
import { pipeline } from 'node:stream/promises';
import { createReadStream, createWriteStream } from 'node:fs';
import { createGzip } from 'node:zlib';

await pipeline(
  createReadStream('input.txt'),
  createGzip(),
  createWriteStream('input.txt.gz')
);

Child Processes

Spawning a shell is a common source of injection vulnerabilities. exec() passes the command through /bin/sh, which means shell metacharacters in user input become code execution. execFile() bypasses the shell entirely.

  • execFile over exec — avoid shell injection. If you need shell features (pipes, globbing), use exec explicitly and sanitize inputs, but prefer execFile by default.
  • Always handle stderr — don't ignore it. At minimum, log it. Silently swallowed stderr hides errors that make debugging impossible later.
  • Set timeout and maxBuffer on exec/execFile — runaway child processes can hang indefinitely or consume unbounded memory.
  • Use signal option with AbortController for cleanup — this lets you cancel child processes cleanly without resorting to kill.
import { execFile } from 'node:child_process';
import { promisify } from 'node:util';

const execFileAsync = promisify(execFile);
const controller = new AbortController();

const { stdout, stderr } = await execFileAsync('git', ['status'], {
  timeout: 10_000,
  maxBuffer: 1024 * 1024,
  signal: controller.signal,
});

Versions & Compatibility

  • Target Node.js 20+ (LTS) minimum unless the project specifies otherwise. Don't use Node.js 18 patterns or polyfills for APIs that landed in v20.
  • Use stable APIs — not experimental — unless explicitly needed. Check the Node.js docs stability index before recommending lesser-known APIs.

package.json

  • "type": "module" for ESM — all new Node.js projects should use ES modules.
  • "engines" field to declare the Node.js version requirement — this gives clear errors when someone tries to run the project on an unsupported version.
  • "exports" field for package entry points — not just "main". The "exports" field supports conditional exports (ESM/CJS), subpath exports, and blocks deep imports into internals.
  • "packageManager" field with corepack for pnpm/yarn version pinning — this ensures everyone on the team uses the same package manager version.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.