Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Mac Slowdown Doctor

    1

    Diagnose and fix macOS lag, memory pressure, and orphaned AI agent processes using native system commands.

    Free

    Security scanned
    Mac Slowdown Doctor

    Mac Slowdown Doctor

    Example session with this skill installed

    My Mac is feeling really laggy and the fans are loud. Can you find out what is going on?

    • Read your context and instructions
    • Compiled the mac slowdown doctor

    I have analyzed your system state and process activity. Your Mac is slow because it is under critical memory pressure, forcing the system to use 4.2 GB of swap on your 8 GB machine.

    Diagnostic Report

    CategoryEvidenceVerdictRecommended Action
    Memory4.2GB Swap, Pressure Level 4CriticalQuit high-memory Electron apps
    CPULoad Avg: 8.42 (4 cores)HighTerminate orphaned node processes
    Disk4.8 GB availableWarningClean npm and Xcode caches
    AI Sessions3 active Claude sessionsNoteClose idle agent CLI instances

    Top Resource Hogs

    1. Docker Desktop: 2.1 GB Memory. Currently holding a large reservation.
    2. Node (PID 84212): 15% CPU. Orphaned process from a claude-code session in ~/projects/api-v2.
    3. Slack: 850 MB Memory. Multiple helper processes active.
    4. Xcode Caches: 12 GB on disk. Found in ~/Library/Developer/Xcode/DerivedData.

    Next steps

    1. Would you like me to gracefully quit Docker Desktop and Slack?
    2. Shall I terminate the orphaned Node process (PID 84212) from your previous agent session?
    3. Should I run npm cache clean --force to recover immediate disk space?

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Your Mac is lagging because background processes and system bottlenecks are fighting for resources. This skill diagnoses exactly why your machine is slow, hot, or loud using native macOS tools. It identifies resource-heavy processes, including those orphaned by AI coding agents, without requiring third-party software.

    What it does

    • System diagnostics runs a snapshot of memory pressure, swap usage, and CPU load to find the primary bottleneck.
    • Process auditing identifies CPU hogs and memory-intensive applications using precise footprint metrics.
    • AI session tracking finds leftover dev servers, watchers, and test runners started by Claude Code, Cursor, or Gemini CLI.
    • Disk analysis scans common developer directories like node_modules, Docker containers, and caches when disk space is critical.
    • Safe remediation provides a ranked list of fixes and only executes kills or cleanups after explicit user approval.

    How it works

    1. Snapshot: The agent runs built-in commands like sysctl, vm_stat, and df to check hardware health.
    2. Investigation: It audits top samples and parent process chains to find the root cause of lag.
    3. Verdict: You receive a plain-language summary of the bottleneck with specific evidence.
    4. Fix: You approve or reject individual actions such as graceful app quits, cache clears, or process termination.

    Frameworks & tools

    This skill relies entirely on native macOS binaries including top, ps, lsof, du, and osascript. It specifically targets processes from AI tools like Claude Code, Cursor, and Gemini.

    Why this beats prompting it yourself

    Generic prompts often rely on the free command or top's first sample, both of which provide misleading data on macOS. This skill uses specific flags to measure compressed memory footprint and thermal throttling symptoms, ensuring you don't kill the wrong process.

    Use cases

    • Identify which AI agent session left a heavy dev server running in the background.
    • Diagnose why kernel_task is spiking and causing thermal throttling.
    • Recover disk space from bloated Library/Developer or Docker volumes.
    • Verify if a machine needs more RAM by analyzing swap usage and memory pressure levels.

    Known limitations

    This skill is strictly for macOS. It cannot terminate system-level processes or those requiring root privileges without manual user intervention.

    Strictly macOS only. Does not use sudo. Cannot kill protected system processes like kernel_task or WindowServer. Requires user approval for all destructive actions.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 10 days ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Listed11 days ago

    What's inside

    Frequently Asked Questions

    This skill specifically targets orphaned processes from AI tools like Claude Code, Cursor, and Gemini CLI, which often leave background dev servers or watchers running after a session ends.

    The skill works entirely with native macOS system binaries like `sysctl` and `vm_stat` and does not require any third-party software or heavy monitoring tools to be installed.

    No, for security reasons the skill only provides recommendations; it will never terminate a process or delete a file without your explicit manual approval.

    While it can identify system-level spikes (like kernel_task), it is restricted to user-level permissions and cannot modify core system files or terminate root processes.

    The skill is optimized for all modern macOS versions and is particularly effective at diagnosing memory pressure on both Intel and Apple Silicon (M1/M2/M3) architectures.

    New in LLM Ops