Mac System Stat

Generate a concise macOS host resource report focused on local-model readiness and machine pressure. Use when the user asks about Mac resource usage, current...

Security Scan

The skill is coherent with its description: it collects local macOS hardware and resource stats by building and running included Python and Swift helpers; it does not request credentials or call external endpoints.

Purpose & Capability

Name/description claim a macOS resource snapshot; the code and runtime instructions only access local system commands (top, ps, vm_stat, system_profiler, sysctl, memory_pressure) and local Swift IOKit/AppleSMC/IOReport helpers — all appropriate for the stated purpose.

Instruction Scope

SKILL.md directs the agent to build and run local Python and Swift helpers that read hardware/OS state. This is within scope, but the helpers compile native binaries and use IOKit/AppleSMC/IOReport (low-level hardware APIs). There are no network calls or attempts to read unrelated credential/config files.

Install Mechanism

No external download/install spec. The skill is source-only and self-builds Swift helpers locally using the system swiftc (Apple Command Line Tools). Built binaries are placed under scripts/bin/ in the skill directory. No remote URLs or package registry pulls are used.

Credentials

The skill requests no environment variables, no credentials, and no config paths. All accessed data are local system interfaces appropriate for a system-status tool.

Persistence & Privilege

always:false (normal). The skill will create compiled helper binaries under its own scripts/bin/ on first run — this writes files to the skill directory but does not modify other skills or global agent config. Helpers run with the invoking user's privileges.

Assessment

This skill appears to do exactly what it says: compile and run local helpers to produce a macOS system snapshot. Before installing, consider: (1) the skill compiles and executes native Swift binaries on your machine using your system's swiftc/Xcode Command Line Tools — review the included Swift sources if you don't fully trust the publisher; (2) the helpers access low-level APIs (IOKit/AppleSMC/IOReport) to read hardware sensors — they run with your user privileges and can read hardware state, but do not perform network exfiltration or request credentials; (3) the skill writes build artifacts to scripts/bin/ within its folder; you may want to run it in a controlled environment or inspect the code prior to first run if you need stronger assurance.

Produce a concise macOS host snapshot using small local helpers.

Use when

• The user asks for current Mac resource usage or machine health
• You need CPU, memory, GPU, swap, power, fan, or a short host summary
• You want a truthful one-shot report, not continuous monitoring

Scripts

• scripts/hoststat — aggregate JSON host snapshot
• scripts/memstat — RAM / compression / swap / memory pressure
• scripts/cpustat — CPU usage, load average, process counts, top CPU processes
• scripts/gpustat — GPU model + live IOAccelerator statistics via local Swift/IOKit helper
• scripts/powerstat — Apple Silicon IOReport-based power sampler (non-privileged, short window average)
• scripts/fanstat — AppleSMC-based fan reader (fan count, RPM, min/max, mode)
• scripts/tempstat — AppleSMC temperature sensors + pmset thermal state (CPU, battery, ambient, board)
• scripts/build-helpers — prebuild Swift helpers explicitly (optional; normal wrappers auto-build as needed)

Layout

• The skill is intentionally self-contained under skills/mac-system-stat/
• Python entrypoints, shared Python modules, Swift sources, and compiled helper output all live under skills/mac-system-stat/scripts/
• scripts/bin/ is generated locally on first run or via scripts/build-helpers

Release contract

• Chosen contract: source-only, self-building on first run
• Ship the skill directory with Python + Swift sources; do not rely on compiled helpers being pre-vendored
• gpustat, powerstat, fanstat, and tempstat will auto-build their helper binaries into scripts/bin/ when missing or stale
• If swiftc / Apple Command Line Tools are unavailable, those four helpers fail truthfully with structured JSON; memstat and cpustat still run normally
• For packaging / ClawHub, exclude transient local artifacts such as scripts/bin/ contents and __pycache__/

Notes

• macOS only; Apple Silicon first
• Default path avoids sudo
• Prefer helper JSON over hand-written summaries
• GPU reads IORegistry properties directly via IOKit instead of parsing ioreg text
• Power uses IOReport Energy Model deltas; values are short-window averages, not hardware-meter absolutes
• Fan RPM comes from AppleSMC F* keys; 0 RPM can mean the fans are currently stopped
• Temperature uses curated AppleSMC keys (Tp0P, Tp0T, Te0T, Ts0P, TB0T, TW0P, Ta0P) + pmset therm state
• Packaging caveat: compiled helpers are not vendored; the target machine still needs Command Line Tools / swiftc for first build unless prebuilt binaries are shipped separately