clean-up

# Safe System Clean-up Skill This skill provides a structured workflow for identifying, auditing, and safely reclaiming disk space across package manager caches, Docker resources, shell backups, and temporary files. > [!IMPORTANT] > **Safety First:** Never execute deletion or purge commands without presenting the targeted files/directories and their sizes to the user and obtaining explicit, written permission. --- ## 1. Audit Phase (Detection) To begin a clean-up, scan the system for space-consuming targets using non-destructive commands. Categorize findings into the following groups: ### A. Package Manager Caches Check the sizes of common package directories: * **uv cache:** `du -sh ~/.cache/uv` * **pip cache:** `du -sh ~/.cache/pip` * **Conda package cache:** `du -sh ~/miniconda3/pkgs` * **npm cache:** `du -sh ~/.npm/_cacache` * **pnpm store:** `du -sh ~/.local/share/pnpm/store` (and check for registered projects via `pnpm store status`) ### B. Docker Storage Check active/stopped containers, unused images, and builder caches: * `docker system df` * `docker ps -a` (to identify any running or stopped containers and their compose project directories) ### C. Transient Agent files & Backups * **Episodic State files:** `.ES/` folders or `ES.md` logs in repository roots. * **Shell history/alias backups:** `~/.bash_history-*.tmp` files, `~/.bash_aliases.bak-*` backups. * **Configuration backups:** Sibling `.bak` files in `~/.openclaw` or `~/.codex`. --- ## 2. Review and C...