Skip to content
cccskills
FreePublished on GitHub

Vitals

Inspect Mac performance to find processes using excessive CPU, GPU, memory, or energy. Combine measurements with explanations of known system processes to diagnose slowness, heat, and startup load, then recommend actions without changing system settings.

OriginalShow the original description

macOS system performance diagnostics — see how the machine is running and what's slowing it down: CPU/GPU/memory/energy hogs, thermal throttling, memory and swap pressure, disk usage, Spotlight indexing, launchd/startup load, via a deterministic read-only CLI with known-process interpretation (kernel_task, WindowServer, mds_stores). USE WHEN mac slow, system slow, what's slowing down my mac, computer is slow, what's eating CPU, CPU usage, GPU usage, what's using the GPU, memory pressure, RAM usage, swap, runaway process, fans loud, mac running hot, thermal throttling, system health, check my mac, how's my system running, top processes, energy hogs, activity monitor, startup items, launch agents load, system taxed. NOT FOR network/wifi diagnostics, website or deployed-app health monitoring, or security scanning.

See how to install

Good for

  • Find the cause of a slow Mac
  • Investigate loud fans and overheating
  • Check memory pressure and swap activity
  • Review system health and startup load

Files included (6)

  • SKILL.md4.8 KB
  • Interpretation.md2.9 KB
  • Tools/Vitals.ts11.8 KB
  • Workflows/DeepDiagnosis.md1.7 KB
  • Workflows/FindCulprit.md1.8 KB
  • Workflows/HealthCheck.md1.2 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Vitals

Read-only macOS performance inspection: a deterministic CLI gathers the numbers, Interpretation.md turns them into a diagnosis instead of a data dump.

Customization

Before executing, check for user customizations at: ~/.claude/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/Vitals/

If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.

Voice Notification

When executing a workflow, do BOTH:

  1. Send voice notification:

    curl -s -X POST http://localhost:31337/notify \
      -H "Content-Type: application/json" \
      -d '{"message": "Running WORKFLOWNAME in Vitals"}' \
      > /dev/null 2>&1 &
    
  2. Output text notification: Running **WorkflowName** in **Vitals**...

Workflow Routing

WorkflowTriggerFile
HealthCheck"how's my system", "check my mac", "system health"Workflows/HealthCheck.md
FindCulprit"mac is slow", "what's eating CPU/GPU/memory", "fans loud"Workflows/FindCulprit.md
DeepDiagnosis"full diagnosis", "deep check", chronic/recurring slownessWorkflows/DeepDiagnosis.md

Interpretation reference (load with any workflow): Interpretation.md — thresholds, known-process table, diagnosis shape.

Quick Reference

bun ~/.claude/skills/Vitals/Tools/Vitals.ts check     # fast snapshot (<1s)
bun ~/.claude/skills/Vitals/Tools/Vitals.ts hogs      # live per-process CPU/energy (~3s)
bun ~/.claude/skills/Vitals/Tools/Vitals.ts full      # everything (~4s)
# also: gpu · memory · disk · thermal · startup · --json · --top N

The tool is read-only by contract: it never kills, renices, unloads, or writes system state. Remediation is recommended to the operator, never executed by the skill.

Examples

Example 1: Something feels slow

User: "My Mac is dragging, what's going on?"
→ FindCulprit: run hogs + check + gpu
→ Cross-reference Interpretation.md (is the hog a known background process?)
→ Report: named culprit with two evidence classes, recommended action for approval

Example 2: Routine check

User: "How's my system doing?"
→ HealthCheck: run check
→ One-screen verdict: 🟢/🟡/🔴 per subsystem, abnormalities flagged against thresholds

Example 3: Fans blasting

User: "Why are my fans so loud?"
→ FindCulprit: thermal + hogs + gpu
→ If kernel_task is the top consumer: that IS the cooling response — report the thermal cause, not the process

Gotchas

  • top's first sample is garbage — CPU% is cumulative since boot. The tool always runs -l 2 and parses the SECOND sample; never "optimize" this away.
  • ps -m does not reliably sort by memory on modern macOS (observed mis-ordering by GB-scale RSS). The tool sorts in code; keep it that way.
  • kernel_task high CPU is the thermal system working, not a runaway process — it pins cores to force cooling. Never recommend killing it.
  • macOS "used" memory is not a problem signal. Free RAM is deliberately spent on cache; kern.memorystatus_vm_pressure_level (1 normal / 2 warning / 4 critical) and live swapouts are the real signals.
  • powermetrics requires sudo and runs forever without -n. It's the only accurate source of per-process energy and GPU power — keep it as the optional consent-gated leg in DeepDiagnosis.
  • top -l truncates command names (~16 chars) regardless of COLUMNS in logging mode. Use the PID to identify the full process (ps -p <pid> -o comm=).
  • GPU via ioreg -c IOAccelerator works without sudo on Apple Silicon (Device Utilization % in PerformanceStatistics) but is not guaranteed on every GPU — the tool degrades to an explicit unavailable-message, never a silent blank.
  • pmset -g therm may not report CPU_Speed_Limit on desktops — the tool treats a missing key as not-throttled rather than erroring.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Compare a component's own design needs with the goals of the larger system through three structured passes. The result identifies conflicts, recommends an appropriate scope, and assesses whether the proposed design fits the whole.

  • Choosing between a service, a library, and inline code
  • Comparing feature requests with product goals
  • Checking how new components fit an existing system
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

Aphorisms

FreeSummary

Find quotes that fit a text’s themes, with attribution and a short reason for each suggestion. Add sourced quotes, research thinkers, and organize a quote collection by theme and usage history to avoid repetition.

  • Finding a fitting quote for a newsletter
  • Saving sourced quotes without duplicates
  • Researching a thinker’s quotes by theme
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

Apify

FreeSummary

Collect social posts, business listings, and product reviews through Apify actors, ready-made data collection tools. Filter results by date, rating, or engagement in code before sending relevant records to the AI for research or lead generation.

  • Find competitors' most popular social posts
  • Build a contact list of local businesses
  • Extract recent negative Amazon reviews
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

Art

FreeSummary

Create blog illustrations, technical diagrams, comparisons, comics, and icons through workflows tailored to each format. The skill guides composition, color, and model selection, then stages generated images for review before they are used in a project.

  • Creating blog header illustrations
  • Explaining a system with a technical diagram
  • Visualizing comparisons and timelines
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

ArXiv

FreeSummary

Find arXiv papers by topic, research category, or paper ID. Retrieve titles, authors, abstracts, and links, check new submissions, and add AlphaXiv's AI-generated overviews when available to help decide which papers to read.

  • Checking new AI and machine learning papers
  • Finding papers on a specific research topic
  • Looking up a paper by its ID or arXiv URL
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

AudioEditor

FreeSummary

Clean spoken recordings by transcribing words, identifying filler and false starts, and joining cuts with short audio fades. Preview proposed edits, reduce breaths, repair noise-gate clicks, and optionally use Cleanvoice for additional cloud processing.

  • Removing filler words and long silences from a podcast
  • Reviewing proposed edits before applying them
  • Repairing clicks caused by a noise gate
danielmiessler/LifeOS19.5KUpdated Sep 4, 2026

See all skills by danielmiessler

Report a problem with this skill