# game-performance-profiler

Turns Unity and Unreal frame-time complaints into one brief: which capture to trust, where the bottleneck sits, what to measure next.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 345 installs · ★ 51
- Author: akillness
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/akillness/jeo-skills/tree/HEAD/.agent-skills/game-performance-profiler/SKILL.md
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/skills/game-performance-profiler

## What it does

Reach for this skill when the evidence is a profiler screenshot, `stat unit` or `stat gpu` output, benchmark-route notes, or a review packet from a target device. It starts by pinning down the context: engine, target platform, whether the build is editor or packaged, the symptom, what evidence exists, how to reproduce it, and quality settings such as resolution and preset. From there it picks a single mode — quick triage, bottleneck classification, benchmark-route planning, device review, or escalation to deeper tools.

Then comes the part that matters: naming one primary bottleneck family. Gameplay scripting on the CPU, draw-call pressure on the render thread, GPU shaders and post-processing, garbage collection and allocation churn, loading and streaming, physics and animation, platform and thermal limits. When the packet is too thin to tell, the answer is `unknown-needs-better-capture`. The skill then recommends the cheapest next capture that would actually separate the candidates: a proper Unity Profiler packet from a player build, a matched `stat unit` and `stat gpu` pair on the exact repro route, or a packaged-on-device capture instead of another editor screenshot. Deeper traces such as Unreal Insights or the GPU Visualizer come only after the first packet earns them.

This is triage and planning, not an automatic optimizer. It does not edit code or tune builds — it hands back one artifact the team can act on.

## Requirements
- Works with: Unity, Unreal Engine

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add akillness/jeo-skills --skill game-performance-profiler -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/game-performance-profiler/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/game-performance-profiler/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add akillness/jeo-skills --skill game-performance-profiler -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/game-performance-profiler/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add akillness/jeo-skills --skill game-performance-profiler -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/game-performance-profiler/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install akillness/jeo-skills game-performance-profiler
```

2. **Manually: put the files in place**
- File: `.github/skills/game-performance-profiler/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add akillness/jeo-skills --skill game-performance-profiler -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/akillness/jeo-skills --path .agent-skills/game-performance-profiler
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/game-performance-profiler/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add akillness/jeo-skills --skill game-performance-profiler -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/akillness/jeo-skills/tree/HEAD/.agent-skills/game-performance-profiler/SKILL.md

## When to use it
- decide which profiler screenshot is worth trusting
- name one primary bottleneck family instead of listing guesses
- pick the next capture that narrows the picture rather than widening it
- tell GPU saturation apart from GC spikes during combat
- judge when to move off quick packets and into Unreal Insights

## Example prompts
- Here is a `stat unit` screenshot from a packaged Steam Deck build; the hitch hits when crossing the market square. Which bottleneck family is this, and what should I capture next?
- Draft a benchmark route for the combat scene: warm-up, repeat count, what to record.
- I have an Unreal Insights trace and an overlay screenshot that disagree. Help me settle on one primary bottleneck family.

## Similar
- [performance-optimization](https://affpapa.org/en/nevibecode/skills/gamedev-skills-performance-optimization.md) — skill · 7700 installs
- [game-development](https://affpapa.org/en/nevibecode/skills/miles990-game-development.md) — skill · 410 installs
- [game-developer](https://affpapa.org/en/nevibecode/agents/game-developer.md) — subagent · 633 installs
- [router](https://affpapa.org/en/nevibecode/skills/router.md) — skill · 6900 installs
- [mashup-mods](https://affpapa.org/en/nevibecode/skills/mashup-mods.md) — skill · 1500 installs
- [reverse-document](https://affpapa.org/en/nevibecode/skills/reverse-document.md) — skill · 461 installs

From the same set: [akillness/jeo-skills](https://affpapa.org/en/nevibecode/by/akillness/jeo-skills.md) — 65 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md