Guides

Straight answers to the questions people ask when a disk fills up. Every one of these is written to be useful whether or not you ever install anything.

• Is it safe to delete node_modules?Yes — as long as the lockfile is intact and you can reinstall. Here is what actually breaks, what does not, and how to check before you delete.
• How to delete node_modules from every project at oncefind, npkill, and a lockfile-verified alternative — the three ways to clear node_modules across a whole machine, and what each one risks.
• Where a developer machine’s disk actually goesDependency trees, build directories and package manager caches, ranked by how much they hold and how cheaply they come back.
• Clearing package manager caches, safelyThe right command for npm, pnpm, yarn, bun, uv, pip, cargo, go, Maven, Gradle, NuGet, vcpkg and Conan — and why you should clear them less often than you think.
• Deleting .venv safelyA virtual environment is disposable by design — but only if the lockfile names everything inside it. How to tell the difference for uv, Poetry, PDM, Pipenv and plain venv.
• Why Rust’s target/ directory gets so bigMultiple profiles, every dependency compiled per project, incremental artefacts and old build fingerprints — where the gigabytes go and which of them are safe to remove.
• dev-prune vs kondo, npkill, cargo-sweep, rimraf and a cron jobAn honest comparison. Four of these are better than dev-prune at the thing they do; here is when each one is the right answer.
• npkill alternatives: beyond the supervised sweepnpkill is good at what it does. What to reach for when you need more than node_modules, a pass that runs itself, or a check stronger than last-modified.
• kondo alternatives: when a prompt is not enoughkondo is the broadest interactive project cleaner there is. What to use when you want the cleanup unattended, verified against lockfiles, or reversible.
• Stress-testing dev-prune: every byte accounted forSix disposable repositories, real installs, one accident. What a logged prune pass deleted, what it refused, and what devp restore put back, byte for byte.
• Where Docker disk space actually goesImages, build cache, stopped containers and volumes each cost you differently to reclaim. What each one is, which are safe to delete, and the one flag that destroys data.
• Giving an AI agent a safe way to free disk spaceAn agent that runs rm -rf has no dry run, no proof and no undo. What to demand of any cleanup an agent performs, and how to teach yours the safer path.
• Cleaning Gradle and Maven build directoriestarget/, build/ and .gradle/ rebuild from the project beside them. The local repository at ~/.m2 does not, and treating it as a cache is how artifacts vanish.