Borrowing it
Nothing to install: this file belongs to ukanwat/aaabench. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ukanwat/aaabench/main/.claude/skills/game-feel/SKILL.mdgit clone --depth 1 https://github.com/ukanwat/aaabenchWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ukanwat/aaabench/game-feel)<a href="https://agentmods.dev/skills/ukanwat/aaabench/game-feel"><img src="https://agentmods.dev/badge/skills/ukanwat/aaabench/game-feel.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00116 | $0.02385 |
| Opus 5 | $0.00058 | $0.01192 |
| Sonnet 5 | $0.00023 | $0.00477 |
| Haiku 4.5 | $0.00012 | $0.00238 |
Grade A, and why
game-feel scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
94% identical to game-feel — 16 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Game feel (juice)
The difference between a mechanic that works and one that feels good is feedback: the layered, slightly-exaggerated response an action provokes. This skill covers the engine- neutral techniques — screen shake, hit-stop, easing, squash & stretch, knockback, and stacked feedback — and tells you how to apply them without burying the underlying simulation. It adds polish on top of an existing mechanic; it does not implement the mechanic.
When to use
- Use when an action (hit, jump, dash, pickup, death, button press) is mechanically correct but feels weak, weightless, or unsatisfying, and you want it to feel responsive and punchy.
- Use to add screen shake, hit-stop/freeze frames, eased motion, squash & stretch, knockback, flashes, or to layer multiple feedback channels onto one event.
- Use to decide how much juice is enough and where it crosses into noise.
When not to use: for the raw controller math (jump height, coyote time) use the
platformer genre and the engine movement skill. For camera follow/deadzone/orbit framing
use camera-systems (this skill only triggers the shake). For mixing, ducking, and adaptive
music use audio-design. For shader-based dissolves/flashes use shader-programming and the
engine shader skill. For the concrete tween/particle node APIs, use the engine animation skill
(godot-animation, unity-animation).
Core principle: feedback is layered and exaggerated
One satisfying hit is usually 5–8 tiny responses firing together within ~100 ms: a sound, a particle burst, a brief hit-stop, a flash, a knockback, a small screen shake, and a number popping up. Each is cheap; stacked, they read as "impact". Two rules keep it from becoming a mess: (1) exaggerate briefly and return to rest (juice is transient, not a new resting state); (2) scale juice to event importance — a footstep is not a boss death.
Core workflow
- Confirm the event hooks exist. Juice attaches to discrete events:
on_hit,on_land,on_pickup,on_death,on_fire. If the mechanic doesn't emit these, add them first. - Pick feedback channels per event from the menu (sound, particles, shake, hit-stop, flash, knockback, tween, number pop). Start with 2–3; add until it reads, then stop.
- Make motion eased, not linear. Route scale/position/UI changes through a tween with an ease (overshoot for "pop", ease-out for "settle"). Linear motion feels robotic.
- Reserve hit-stop and shake for impact. They are the strongest, most abusable tools — short durations, scaled to importance, and never on routine actions.
- Keep feedback off the critical simulation. Shake moves the camera/visual, not the body; hit-stop uses time scale or a real-time pause, not a gameplay-logic stall.
- Tune by importance tiers. Define small/medium/large feedback presets and assign events to a tier, so the whole game's juice stays consistent and proportional.
- Verify by playing and watching. Trigger the event repeatedly; confirm the feedback fires, returns to rest, and is not nauseating or input-blocking. Report what you observed (does shake decay? does input still register during hit-stop?).
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 172 lines · 116 tokens per session scan A 89dc36afaeb4
game-feel is a skill published in the GitHub repository ukanwat/aaabench (382 stars, last pushed 24d ago), licensed MIT. It adds 116 tokens to every session and 2,385 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. It is 94% identical to game-feel, differing in 16 lines, and is treated as a copy.
Other skills, from other repositories
unreal-engineer
!cat skills/shared/protocols/3d-spatial-foundations.md 2>/dev/null || echo "=== 3D Foundations not loaded ===".
unreal-playtest-agent
Domain skill - run screenshot-light PIE playtest episodes with structured entity observations, bounded semantic actions, transition polling, and in-memory traces for QA and external policy or RL runners.
unreal-fab-assets
Acquire free Fab marketplace content through Unreal Engine's official Fab integration, then verify the imported Content Browser assets. Use when the user asks to find, download, add, or import Fab or Unreal Marketplace assets. Not for arbitrary file imports - use unreal-assets instead.
unreal-pie
Domain skill - Unreal Engine Play-In-Editor (PIE) closed-loop verification. Provides PIE lifecycle control (enter/pause/resume/exit), controlled input injection (no OS dependency), viewport screenshot capture, output log snapshot, performance sampling, and Automation Test execution with async job polling. Use for…
unreal-actors
Domain skill - Unreal Engine actor management: list, spawn, transform, and delete level actors. Use when inspecting or editing actors in the current Unreal Editor level. Not for Content Browser asset import/export - use unreal-assets for that.
unreal-automation
Domain skill - Unreal Engine native Automation Test and MCP health validation. Use to inspect typed plugin readiness, list UE Automation tests, queue native test runs, and run a safe self-check against the active MCP server.