About this skill

Problem

A slow computer is rarely caused by one issue: cache buildup, bloated build artifacts, excessive startup items, and heavyweight toolchains make triage slow. Deleting files, editing the registry, or killing processes without context can damage user data or system stability.

How It Works

The skill constrains cleanup into reviewable engineering steps:
- System check first: collect information such as sw_vers, systeminfo, /etc/os-release, disk usage, and top-level large files before acting.
- Risk-graded execution: classify actions as A safe, B moderate, or C sensitive. npm/pip caches and temp directories can be cleaned; node_modules and DerivedData require impact notes; startup items, services, registry changes, and system processes require explicit confirmation.
- Record and script: log each command, rationale, and risk level, then generate a repeatable PC清理-一键.command, PC清理-一键.bat, or PC清理-一键.sh script.

Boundaries

It is useful for routine disk cleanup, startup-item review, and tool replacement advice, but it does not replace a full IT operations process. It avoids user data directories such as ~/Documents, photo libraries, and chat logs, and it does not write one-off dangerous operations into repeatable scripts.

Screenshots

Use Cases

When a dev workstation's C drive or root filesystem is nearly full, collect disk and large-file data before clearing regenerable caches and build artifacts.
When macOS feels slow or has too many background items, review startup items and services item by item, confirming C-level changes only after user consent.
When a Linux development environment needs repeatable cleanup, turn safe steps into a one-click desktop script instead of manual commands every time.
When a toolchain is bloated, compare current tools with replacements by workload, including migration cost and configuration steps.

Best For

Engineers managing multiple dev machines who need risk-graded disk and startup-item reviews plus repeatable cleanup scripts.
macOS workstation maintainers who want to clean caches and build artifacts while confirming service or proxy changes item by item.
Linux engineers who need installed-tool audits, clear replacement recommendations, and configuration steps.
IT support staff who need cleanup commands, impact notes, and risk levels recorded as reviewable procedures.