home / skills / serejaris / ris-claude-code / macos-fixer

macos-fixer skill

safe

This skill diagnoses macOS memory and performance issues and provides actionable fixes to improve responsiveness and reduce memory pressure.

npx playbooks add skill serejaris/ris-claude-code --skill macos-fixer

Review the files below or copy the command above to add this skill to your agents.

Files (1)

skill.md

3.1 KB

---
name: macos-fixer
description: Use when system feels sluggish, apps unresponsive, fans spinning, or user asks about memory/RAM/performance - diagnoses macOS memory and suggests fixes
---

# macOS Fixer

## Diagnostic Commands

Run in order, stop when problem found:

```bash
# 1. Quick overview (RAM, swap, pressure)
top -l 1 -s 0 | head -12

# 2. Memory pressure percentage
memory_pressure | grep "free percentage"

# 3. Swap details
sysctl vm.swapusage

# 4. Top memory consumers (MB)
ps -eo pid,rss,comm | sort -k2 -rn | head -15 | awk '{printf "%6d MB  %s\n", $2/1024, $3}'

# 5. VM statistics (compressions, pageouts)
vm_stat | grep -E "(Compressions|Decompressions|Swapins|Swapouts|compressor)"

# 6. Dev servers (common leaks)
lsof -i -P 2>/dev/null | grep node | grep LISTEN

# 7. Chrome tabs proxy
pgrep -l "Google Chrome" | wc -l
```

## Decision Tree

```
Memory pressure < 30%? ──yes──→ CRITICAL: close apps NOW
     │no
     ▼
Swap > 30% of RAM? ─────yes──→ RESTART RECOMMENDED
     │no
     ▼
High swap I/O? ─────────yes──→ MEMORY THRASHING: find the leak
     │no
     ▼
Uptime > 5 days? ───────yes──→ SCHEDULE RESTART
     │no
     ▼
Node process > 700MB? ──yes──→ Kill dev servers
     │no
     ▼
Chrome > 40 processes? ─yes──→ Close tabs or restart Chrome
     │no
     ▼
SYSTEM OK
```

## Quick Fixes

| Problem | Command |
|---------|---------|
| Kill Vite | `pkill -f "vite"` |
| Kill Next.js | `pkill -f "next-server"` |
| Clear FS cache | `sudo purge` |
| Restart Chrome | `pkill -x "Google Chrome" && open -a "Google Chrome"` |
| Restart Finder | `killall Finder` |
| Full restart | `sudo shutdown -r now` |

## Common Memory Hogs

| App | Typical | Leak Signs |
|-----|---------|------------|
| node (dev) | 200-400 MB | > 700 MB, multiple instances |
| Chrome | 50-100 MB/tab | > 40 processes |
| Electron apps | 200-400 MB | multiple renderer processes |
| VSCode | 300-600 MB | extensions bloat |

## Report Format

```
## Memory Status: [OK/WARNING/CRITICAL]

| Metric | Value | Status |
|--------|-------|--------|
| RAM used | X/Y GB | |
| Swap | X GB | |
| Pressure | X% free | |
| Uptime | X days | |

**Top Consumers:**
1. app: X MB
2. app: X MB
3. app: X MB

**Issues:** [list or "none"]
**Actions:** [numbered list or "none needed"]
```

## Misconceptions

| Myth | Reality |
|------|---------|
| "High used RAM = problem" | macOS caches aggressively; check pressure not usage |
| "Memory cleaner apps help" | They force cache purge, causing MORE disk I/O |
| "Swap = always bad" | Swap with green pressure = fine, system optimizing |
| "Compressor = problem" | Compressor is GOOD — saves RAM without disk I/O |
| "Free RAM should be high" | Unused RAM is wasted RAM; macOS uses it for cache |

## Preventive Measures

1. **Restart schedule**: Every 3-5 days, or when swap > 30% of RAM
2. **Dev cleanup**: Kill node servers after work session
3. **Chrome hygiene**: Use tab suspender extension, limit tabs
4. **Monitor pattern**: High swap I/O without high swap usage = leak resolved but residue remains → restart soon

Overview

This skill diagnoses macOS memory and performance issues and suggests practical fixes when the system feels sluggish, apps are unresponsive, or fans spin up. It runs a small set of shell checks to report RAM, swap, memory pressure, top consumers, and common developer-server problems, then recommends targeted actions like killing runaway processes or scheduling a restart.

How this skill works

It runs a sequence of lightweight macOS commands to gather RAM usage, memory pressure, swap usage, VM stats (compressions/pageouts), and the biggest memory consumers. The skill applies a simple decision tree to classify the state (OK, WARNING, CRITICAL) and maps conditions to concrete fixes such as killing specific dev servers, restarting apps, or rebooting. It also highlights preventive steps to reduce recurrence.

When to use it

System feels sluggish or apps are slow to respond
Fans run loudly or system temperature rises
You suspect memory pressure, swapping, or leaks
After long uptime (>3–5 days) or heavy dev sessions
When Chrome or node processes consume lots of RAM

Best practices

Check memory pressure, not raw used RAM — pressure reveals real stress
Run diagnostics in order and stop when a clear issue is found
Prefer targeted process kills (vite, next, node) before a full restart
Avoid memory-cleaner apps; they purge useful cache and increase disk I/O
Schedule periodic restarts (every 3–5 days) if you run many dev services

Example use cases

Diagnose high swap usage and decide whether to restart or hunt a leak
Identify a runaway node process and kill the dev server with a single command
Detect excessive Chrome renderer processes and suggest closing tabs or restarting Chrome
Confirm compressor activity vs true swapping, avoiding unnecessary panic-reboots
Provide a quick report (status, top consumers, actions) to share with support or teammates

FAQ

When is restart required?

If swap exceeds ~30% of RAM, or memory pressure shows heavy swapping/io, a restart is recommended to clear residue and stop thrashing.

Is high RAM usage always bad?

No. macOS uses RAM for cache. Check memory pressure and swap activity — green pressure with high used RAM is normal.