clean-system
Skill8 starsChanged 9 days agoMITView on GitHub
Reclaim disk space or clear the clutter Claude Code, Codex and Elpis leave (old chat logs, file history, logs, backups), or name or remove single sessions. Agent clutter goes without asking; caches and Docker need a yes.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- 0xrizz/os-managertheir own version, first seen Aug 28, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- 0xrizz/os-manager★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository17 other files it gives its agents.
- app-controlSkill
- artifactSkill
- contribution-checklistSkill
- debuggingSkill
- doc-styleSkill
- experiment-workflowSkill
- first-principlesSkill
- latex-authoringSkill
- logoSkill
- multi-agentSkill
- product-designSkill
- project-docsSkill
- reference-driven-productSkill
- specSkill
- tddSkill
- test-auditSkill
- word-farsi-docxSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Run python3 clean.py from this folder (dry run), show the per-category totals, then run it with --apply.
- Very specific: line 12
One-off targets: add them with --also PATH ....
- Very specific: line 18
If an auto-mode classifier blocks --apply, don't work around it.
- Very specific: line 29
Docker: run docker system df. Images, containers and build cache are separate approvals.
- Distinctive: line 30
Never target a home directory, workspace root, unresolved variable or broad glob.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
System clean-up
1. Agent clutter: clear it without asking (rule 27)
Run python3 clean.py from this folder (dry run), show the per-category totals, then run it with --apply.
- What it clears: old Claude, Codex and Elpis chat logs, Claude file history and scratch, Codex and Elpis logs, Codex per-run configs, and backups (
*.bak,*.pre-*,*.before-*,*backup*) in the agent config folders. - One-off targets: add them with
--also PATH .... - What it always spares:
- anything touched in the last 24 hours, or open in a running process;
- memory folders, credentials, live
config.tomlandsettings.json, databases, prompt history; - skills, plugins, jobs, plans,
ES.md.
- Everything goes to the trash (
gio trash). - If an auto-mode classifier blocks
--apply, don't work around it. Give the user the exact! python3 … --applycommand to run.
2. Everything else: audit, then ask
Package caches and Docker hold other projects' state, so list the targets with sizes and get the user's yes per category.
- Caches:
- uv:
du -sh ~/.cache/uv, cleaned byuv cache clean. - pip:
~/.cache/pip, cleaned bypip cache purge. - conda:
~/miniconda3/pkgs, cleaned byconda clean --all -y. - npm:
~/.npm/_cacache, cleaned bynpm cache clean --force. - pnpm:
pnpm store prune. Never delete the store folder.
- uv:
- Docker: run
docker system df. Images, containers and build cache are separate approvals. Volumes need their own explicit yes because they may hold irreplaceable data. - Never target a home directory, workspace root, unresolved variable or broad glob.
3. Verify
Rerun the dry run and df -h /. Report what moved and how much space was freed. Note that it all sits in the trash until the user empties it.
4. Single sessions: name or remove one at a time
For a recent session the user wants named or judged individually, read only its product's own history root (Claude Code ~/.claude/projects/, Codex ~/.codex/sessions/, Elpis ~/.elpis-next/sessions/), list id, short summary, last activity and message count in one table, and propose a 3 to 6 word name from the first meaningful user message. Rename or remove only through the product's own command, never by editing raw logs or indexes, and never the current session or one with a live writer.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.