Skills for Claude Code and other AI Agents

[rudolfolah]

A set of skills that can be used by Claude Code and other AI agents. To ensure that most of the skills are usable, I used Oh My Pi with GPT 5.5 and GPT 5.6 Sol to generate programs and outfit them with performance profiling tools, only tested on a MacBook Pro. If the skills do not work, please create an issue.

Skills:

• Browser JavaScript Profiling
• JavaScript Profiling (Node.js)
• GPU Profiling
• Game Profiling: Godot and Unity
• JVM Profiling
• Python Profiling
• Python CPU Profiling: cProfile, py-spy, yappi, pyinstrument
• Python DTrace Profiling
• Python Memory Profiling
• Systems Profiling: gprof, Tracy, Valgrind, Apple Instruments

[rudolfolah]

The skill needs to be independent and should not reference other skills.

Suggested change

This is **browser-page profiling**, not Node.js process profiling. The existing JavaScript/Node.js skill covers `node --cpu-prof`, `node --heap-prof`, `--trace-gc`, V8 heap APIs, and server-side tools. Do not use those Node flags as a substitute for recording a page: they profile a Node process and its isolate. A Node `.cpuprofile` or `.heapprofile` can be opened in DevTools, but it is not a browser Performance trace and does not include the page's renderer, network, layout, paint, or user interaction. Use the Performance and Memory panels below for a browser page.

This is **browser-page profiling**, not Node.js process profiling. Node.js flags include `node --cpu-prof`, `node --heap-prof`, `--trace-gc`, V8 heap APIs, and server-side tools. Do not use those Node flags as a substitute for recording a page: they profile a Node process and its isolate. A Node `.cpuprofile` or `.heapprofile` can be opened in DevTools, but it is not a browser Performance trace and does not include the page's renderer, network, layout, paint, or user interaction. Use the Performance and Memory panels below for a browser page.

[rudolfolah]

A skill could be added for profiling React apps.

[rudolfolah]

There are MCPs for Unity and Godot, those can be referenced:

Use the Unity MCP if available.

[rudolfolah]

This can be split into supporting files, for NVIDIA, AMD, Android, Intel, Metal.

[rudolfolah]

Deno and Bun have support for --cpu-prof

[rudolfolah]

All the references to Python 3.11 and .python-version should be removed. DTrace has been supported since Python 3.6.

[rudolfolah]

Suggested change

	Use this skill when investigating CPU time, wall-clock latency, Python allocations, process memory, object growth, or interpreter-level events in the examples under `python/`. The repository compares multiple profilers around `python/program.py`; choose the narrowest tool that can answer the question instead of running every profiler by default.
	Use this skill when investigating CPU time, wall-clock latency, Python allocations, process memory, object growth, or interpreter-level events. Choose the narrowest tool that can answer the question instead of running every profiler by default.

[rudolfolah]

Suggested change

	The examples target the Python version in `python/.python-version` and use the virtual environment described in `python/README.md`:

[rudolfolah]

Suggested change

	`program.py` fetches a Wikipedia page over HTTPS before parsing it. Treat a non-empty top-ten word list as the baseline preflight; `[]` means the intended parsing workload did not run and the resulting profile is not useful. The current target calls `requests.get(...)` without a descriptive `User-Agent` or `raise_for_status()`, so Wikimedia may return its small HTTP 403 robot-policy response and the program may silently print `[]`. If that happens, stop before profiling and make the workload valid: preferably use a checked-in/local fixture, or update the request to use a policy-compliant descriptive `User-Agent` and fail on non-success responses. Do not compare a rejection-response profile with a successful-page profile.
	Even with a successful response, network, DNS, TLS, server, and response-size variance can dominate timings. For repeatable comparisons, keep the input and environment fixed; never treat one network run as a reliable regression measurement.
	The repository's comparison runner is:
	```bash
	./timing.sh
	```
	It invokes cProfile, tracemalloc, Memray, pyinstrument, yappi, Filprofiler, psutil, and guppy3. Profiling changes both runtime and memory behavior, so compare profiles only under the same Python build, dependency versions, input, warm-up state, and command-line options.
	Profiling changes both runtime and memory behavior, so compare profiles only under the same Python build, dependency versions, input, warm-up state, and command-line options.

[rudolfolah]

Split into supporting files.