What is Dependency? Definition & Meaning

A dependency is a third-party package or library that your project relies on to function. When AI generates code that uses Stripe, Prisma, or Tailwind, those are dependencies — external code your project needs. Managing dependencies wisely prevents bloat, security vulnerabilities, and maintenance headaches.

Dependencies are a double-edged sword. They save you from reinventing the wheel but also introduce code you don't control into your project.

Direct vs Transitive Dependencies

Direct — Packages you explicitly install (listed in your package.json)
Transitive — Packages your dependencies depend on (installed automatically)

A project with 25 direct dependencies might have 500+ transitive dependencies. That's a lot of external code.

The Dependency Tradeoff

Benefit	Risk
Save development time	More code to maintain
Battle-tested solutions	Potential security vulnerabilities
Community support	Can become abandoned
Proven patterns	Version conflicts

AI and Dependencies

AI sometimes adds unnecessary dependencies. Watch for:

Overlapping packages — Two libraries that do the same thing
Heavy dependencies for simple tasks — A 100KB library for one small function
Outdated packages — AI might suggest older, deprecated versions
Unnecessary additions — Built-in solutions exist but AI reaches for a library

Best Practices

Review what AI installs — Do you actually need each package?
Prefer popular packages — More users means more maintenance and security fixes
Keep dependencies updated — Security patches matter
Audit regularly — npm audit checks for known vulnerabilities
Use lock files — Ensure consistent versions across environments

When to Add a Dependency

Ask: "Is this complex enough to warrant external code?" If you can write it in 20 lines, maybe you don't need a package for it.

Dependency

Example

Direct vs Transitive Dependencies

The Dependency Tradeoff

AI and Dependencies

Best Practices

When to Add a Dependency