cleanup

name	cleanup
description	Clean workspace — build artifacts, dependencies, caches, Docker volumes, or temp files. Accepts arguments like build, deps, all, cache, docker, tasks, reinstall.

Cleanup Skill

Clean up workspace artifacts based on the requested scope.

Usage

$cleanup              # Default: show what can be cleaned
$cleanup build        # Remove build artifacts
$cleanup deps         # Remove dependency directories (node_modules, venv, etc.)
$cleanup all          # Full clean: deps + build artifacts
$cleanup cache        # Clear build tool caches
$cleanup docker       # docker system prune -f (no named volumes removed)
$cleanup tasks        # Remove obsolete task files (lessons.md is NEVER deleted)
$cleanup reinstall    # Full reinstall: clean deps → install

Instructions

Determine action from $ARGUMENTS:

Argument	Action
build or dist	Remove all build artifacts (dist/, build/, .next/, target/)
deps	Remove dependency directories (node_modules/, venv/, .venv/, target/)
all	Full clean: deps + build artifacts, then reinstall
cache	Clear build tool caches (.turbo/, .cache/, __pycache__/, .mypy_cache/)
docker	docker system prune -f (does NOT remove named volumes)
tasks	Clean obsolete task files (see rules below)
reinstall	Full reinstall: clean deps → install

Tasks file cleanup rules:

• NEVER delete brain/tasks/lessons.md (accumulated wisdom)
• NEVER delete brain/tasks/CODEX_ERRORS.md (bug history)
• brain/tasks/todo.md can be archived if no active task
• brain/tasks/mr-description-*.md can be deleted after MR is merged
• brain/tasks/ticket-*.md can be deleted after ticket is created

Safety rules:

• NEVER run rm -rf /
• NEVER delete .env or .env.* files
• For all: confirm before deleting dependencies if package.json/pyproject.toml shows install script hooks
• For docker: confirm before running prune in production-adjacent environments