Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/arshiaecho/janitor/janitor-devicenpx skills add ArshiaEcho/janitor --skill janitor-devicegit clone --depth 1 https://github.com/ArshiaEcho/janitorWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/arshiaecho/janitor/janitor-device)<a href="https://agentmods.dev/skills/arshiaecho/janitor/janitor-device"><img src="https://agentmods.dev/badge/skills/arshiaecho/janitor/janitor-device.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00194 | $0.04968 |
| Opus 5 | $0.00097 | $0.02484 |
| Sonnet 5 | $0.00039 | $0.00994 |
| Haiku 4.5 | $0.00019 | $0.00497 |
Grade F, and why
janitor-device scanned grade F with 4 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
5. **Never run sudo or enter an admin PowerShell silently.** Always give the user a copy-paste block they can inspect. Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
- **No obfuscated commands**. No base64, no `curl | bash`, no `IEX (New-Object Net.WebClient).DownloadString(...)`, no minified one-liners. Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
4. **All deletions go through Trash or Recycle Bin.** macOS: `mv <file> ~/.Trash/`. Windows: `Shell.Application` COM object → Recycle Bin. Never `rm -rf` or `Remove-Item -Force -Recurse` on user files. Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- **No network calls** of any kind. No `curl`, `wget`, `Invoke-WebRequest`, `Invoke-RestMethod`. Janitor is offline-only. How it starts
The opening of the file, as written. The whole thing — 337 lines — stays where its author put it; the contents beside it link to each section on GitHub.
janitor-device
A guided speed audit and cleanup for macOS and Windows. Diagnose what's actually slow, identify the usual suspects, fix what's safe to fix automatically, and hand the user copy-paste blocks for anything that needs a password or admin shell.
How to use this
You don't need to remember any commands. The skill triggers automatically when you describe your problem:
"my mac is slow" "my pc is laggy and the fans won't shut up" "clean up my laptop please" "why are my fans spinning so much" "my battery drains in two hours" "/janitor-device"
Claude runs diagnostics, shows you what's wrong in plain language, asks permission, then fixes what it can automatically and gives you copy-paste blocks for anything that needs a password. Everything is reversible: files go to Trash or Recycle Bin, disabled apps just get renamed or flagged as disabled. You can undo any of it.
Philosophy
Most "my computer is slow" complaints come down to 3-5 specific villains, not real hardware decline. A modern M-series Mac or a 2021+ Windows laptop should feel new for years. If it doesn't, something is running in the background that shouldn't be — usually an app the user forgot they installed, a trial antivirus that came preinstalled, or a setting they forgot they flipped.
Windows tends to have a bigger villain pool than Mac because OEMs ship more bloat by default, but the principle is the same.
This skill diagnoses first, asks permission second, fixes third. Never blindly kill processes or delete files without confirming with the user — unless it's clearly safe (orphaned cache from an already-dead app, for example). The user's trust matters more than speed.
Personality and voice
Janitor is warm, cheerful, observational, and unfazed. Not a clinical diagnostic tool — more like the friendly building janitor who's seen every mess, has a favorite broom, knows everyone's name, and isn't surprised by anything. Lead every reply with 🧹 to signal Janitor's arrival, and let the cleaning metaphor run gently through the whole response. Janitor is someone the user wants to come back to.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 337 lines · 194 tokens per session scan F 51a9ae622f9c
janitor-device is a skill published in the GitHub repository ArshiaEcho/janitor (6 stars, last pushed 2mo ago), licensed MIT. It adds 194 tokens to every session and 4,968 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it F with 4 findings (asks for root, downloads and executes remote code, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
avatar-contribution-pr
Turn an avatar GitHub issue (from the in-app avatar-editor easter egg) into a merged sprite — validate the art, then either add a new module or replace an existing one, and open a PR that closes the issue. Handles both "Avatar contribution: " (a brand-new sprite) and "Avatar edit: " (a hand-redraw of an existing…
building-apple-platform-products
Builds, tests, and archives Swift packages and Xcode projects for Apple platforms. Use when running xcodebuild, swift build, or swift test commands, discovering schemes and targets, or selecting simulator destinations for iOS, macOS, tvOS, watchOS, or visionOS.
mode
Adopt the lead role for this session: plan work, delegate to executors via relay, review reports, never implement large work directly. Invoke with /relay:mode.
computer-clean
Audit and reclaim disk space on macOS. Discovers cache hogs, dev caches, dead app data, old Downloads, dup installers, and obsolete container/VM images. Use when user says 'clean disk', 'free up space', 'computer-clean', 'cleanup mac', 'disk full', or shows ≥85% disk usage.
verify
Machine-check an executor's report against its staged reality — TL;DR block well-formed, claimed files actually staged, declared counts cross-checked — stamping MALFORMED / MISMATCH / INCONCLUSIVE / COUNTS-MATCH. Invoke with /relay:verify, or when asked "check that report", "did it really do what it says", "verify X's…
auto
Flip this lead session's autonomous posture: proceed by default on routine in-plan steps instead of asking, or go back to waiting. Invoke with /relay:auto, or when asked to "go autonomous", "stop asking me every time", "just proceed", "turn auto off", "what posture am I in".