agentleFS
d-padmanabhand-padmanabhan/agent-engineering-handbook/skills/typescript-javascript/SKILL.md

typescript-javascript

Skill17 starsChanged 55 days agoMITView on GitHub

TypeScript and JavaScript development standards for modern web and Node.js development. Covers strict TypeScript configuration, type safety patterns, ESM modules, async/await, testing with Jest/Vitest, and security best practices. Use when working with .ts, .tsx, .js, .mjs files, package.json, tsconfig.json, or when asking about TypeScript/JavaScript best practices.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 5 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Explicit Over Implicit: Prefer explicit types for clarity and maintainability

    line 13
  2. Very specific:

    Security First: Never use eval, sanitize HTML, validate inputs

    line 15
  3. Very specific:

    Every service/network fetch must include a timeout, check response.ok, and validate the JSON shape before returning typed data.

    line 42
  4. Very specific:

    Use the observability skill (${HANDBOOK_ROOT}/skills/observability/SKILL.md) for event schemas, redaction, metric cardinality, trace …

    line 68
  5. Distinctive:

    Keep runtime-specific logger and OpenTelemetry integration in this skill.

    line 71
  6. Distinctive:

    Add TSDoc or JSDoc when callers need behavior, error, ownership, blocking, side-effect, constraint, or lifecycle information that the …

    line 90
  7. Very specific:

    Use @param, @returns, @throws, @remarks, and examples only when they add contract information.

    line 91
  8. Distinctive:

    Do not add comments that merely repeat names or types.

    line 92
  9. Very specific:

    Never use eval, new Function, or unsanitized innerHTML

    line 229
  10. Very specific:

    Use textContent for DOM insertion

    line 230
  11. Very specific:

    Do not log secrets/tokens/PII

    line 232
  12. Distinctive:

    Use parameterized queries.

    line 233

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

TypeScript & JavaScript Development

Mandatory gates are owned by the JavaScript and TypeScript rule (${HANDBOOK_ROOT}/rules/225-javascript-typescript.mdc). This skill and its references own procedures and examples.

Guiding Principles

  1. Type Safety: Leverage strict mode, avoid any, use discriminated unions
  2. Explicit Over Implicit: Prefer explicit types for clarity and maintainability
  3. Modern Defaults: ESM, const/let, async/await, optional chaining
  4. Security First: Never use eval, sanitize HTML, validate inputs

Quick Reference

AspectTypeScriptJavaScript
Package Managerpnpm preferredpnpm preferred
Module SystemES ModulesES Modules + // @ts-check
Lintingeslint --max-warnings=0eslint --max-warnings=0
FormattingPrettierPrettier
TypesStrict modeJSDoc types

Non-Negotiables

NN-1: Validate at trust boundaries

Static types do not validate runtime data. Treat HTTP bodies, responses, webhooks, queue messages, localStorage/sessionStorage, environment variables, CLI args, and JSON files as unknown until validated with Zod or an explicit type guard.

Reject:

  • JSON.parse(raw) as T
  • await response.json() as T
  • unchecked property access on remote JSON
  • as any or double assertions (value as unknown as T)

NN-2: Fetch/API calls require timeout + status check + schema validation

Every service/network fetch must include a timeout, check response.ok, and validate the JSON shape before returning typed data.

import { z } from 'zod';

async function fetchJson<T>(
  url: string,
  schema: z.ZodType<T>,
  options: RequestInit = {},
): Promise<T> {
  const response = await fetch(url, {
    ...options,
    signal: options.signal ?? AbortSignal.timeout(5_000),
    headers: { Accept: 'application/json', ...options.headers },
  });
  if (!response.ok) {
    throw new Error(`request failed: HTTP ${response.status}`);
  }
  return schema.parse(await response.json());
}

NN-3: Production servers are hardened by default

Node/Express/Fastify services need bounded request bodies, explicit CORS, security headers, request IDs, structured logging, generic 5xx responses, and graceful shutdown. Do not use direct app.listen(...) examples for production services without retaining and closing the server.

Use the observability skill (${HANDBOOK_ROOT}/skills/observability/SKILL.md) for event schemas, redaction, metric cardinality, trace propagation, SLOs, alerts, and telemetry pipelines. Keep runtime-specific logger and OpenTelemetry integration in this skill.

NN-4: Package manager and runtime are pinned

Declare packageManager and Node engine policy in package.json; CI uses Corepack and the declared package manager.

Production applications normally use the newest supported active LTS Node.js line accepted by the deployment target. Pin the exact package manager and tested CI runtime, commit one native lockfile, and test compiler upgrades. Follow the dependency and toolchain currency workflow (${HANDBOOK_ROOT}/skills/core-engineering/references/dependency-and-toolchain-currency.md).

{
  "packageManager": "pnpm@10.0.0",
  "engines": { "node": ">=22" }
}

NN-5: Exported APIs have explicit type and documentation contracts

  • Production JavaScript uses // @ts-check and precise JSDoc types for exported parameters, return values, callbacks, generics, and shared object shapes.
  • TypeScript exported functions and public methods use explicit parameter and return types. Keep type inference for obvious local variables.
  • Add TSDoc or JSDoc when callers need behavior, error, ownership, blocking, side-effect, constraint, or lifecycle information that the signature cannot express.
  • Use @param, @returns, @throws, @remarks, and examples only when they add contract information.
  • Do not add comments that merely repeat names or types.
  • Static types and documentation never replace runtime validation at trust boundaries.

Critical Patterns

// 1. Use strict equality
if (value === 0) { }        // ✅ GOOD
if (value == 0) { }         // ❌ BAD

// 2. Handle Promise rejections
fetchData().catch(err => console.error(err));  // ✅ GOOD

// 3. Optional chaining and nullish coalescing
const name = user?.profile?.name ?? 'Guest';  // ✅ GOOD

// 4. Use Set/Map for lookups
const seen = new Set();     // ✅ GOOD (O(1))
const seen = [];            // ❌ BAD (O(n))

// 5. Never use eval or new Function
eval(userInput);            // ❌ NEVER DO THIS

// 6. Sanitize HTML
element.textContent = userInput;  // ✅ GOOD
element.innerHTML = userInput;    // ❌ BAD (XSS)

TypeScript Configuration

{
  "compilerOptions": {
    "target": "ES2022",
    "module": "ESNext",
    "moduleResolution": "bundler",
    "strict": true,
    "noImplicitAny": true,
    "strictNullChecks": true,
    "noUnusedLocals": true,
    "noUnusedParameters": true,
    "noUncheckedIndexedAccess": true,
    "esModuleInterop": true,
    "skipLibCheck": false
  }
}

Avoid any - Use Proper Types

// ❌ BAD - Loses all type safety
function processData(data: any): any {
  return data.value;
}

// ✅ GOOD - Generic type
function processData<T>(data: T): T {
  return data;
}

// ✅ GOOD - Unknown for truly unknown types
function processData(data: unknown): string {
  if (typeof data === 'object' && data !== null && 'value' in data) {
    return String((data as { value: unknown }).value);
  }
  throw new Error('Invalid data');
}

Discriminated Unions

type SuccessResponse = {
  status: 'success';
  data: { id: string; name: string };
};

type ErrorResponse = {
  status: 'error';
  error: { code: number; message: string };
};

type ApiResponse = SuccessResponse | ErrorResponse;

function handleResponse(response: ApiResponse): void {
  if (response.status === 'success') {
    console.log(response.data.id);  // Type-safe
  } else {
    console.log(response.error.message);  // Type-safe
  }
}

Utility Types

interface User {
  id: string;
  name: string;
  email: string;
  password: string;
}

type UserUpdate = Partial<User>;           // All optional
type UserCredentials = Pick<User, 'email' | 'password'>;
type UserPublic = Omit<User, 'password'>;  // Exclude password
type RequiredUser = Required<User>;        // All required
type ReadonlyUser = Readonly<User>;        // Immutable

Async Patterns

// Parallel execution
const [users, products] = await Promise.all([
  fetchUsers(),
  fetchProducts()
]);

// Handle partial failures
const results = await Promise.allSettled([
  fetchUsers(),
  fetchProducts()
]);

results.forEach(result => {
  if (result.status === 'fulfilled') {
    console.log(result.value);
  } else {
    console.error(result.reason);
  }
});

Security Rules (Mandatory)

  • Never use eval, new Function, or unsanitized innerHTML
  • Use textContent for DOM insertion
  • Validate and sanitize all external inputs
  • Do not log secrets/tokens/PII
  • Use parameterized queries; no string-built queries
  • Enforce HTTPS; secure cookies (HttpOnly, SameSite)

Naming Conventions

TypeConventionExample
Functions/VariablescamelCasefetchUserData
Classes/InterfacesPascalCaseUserService
ConstantsUPPER_SNAKE_CASEMAX_RETRIES
TypesPascalCaseApiResponse

Detailed References

  • TypeScript Patterns: See TypeScript patterns (${HANDBOOK_ROOT}/skills/typescript-javascript/references/typescript-patterns.md) for advanced types, generics, and mapped types
  • JavaScript Patterns: See JavaScript patterns (${HANDBOOK_ROOT}/skills/typescript-javascript/references/javascript-patterns.md) for JSDoc, ESM, and performance

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.

Reviews can't be read right now.

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.