Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/thearcforge/hades/unity-performancenpx skills add TheArcForge/Hades --skill unity-performancegit clone --depth 1 https://github.com/TheArcForge/HadesWhat 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 | $0.00032 | $0.06734 |
| Opus 5 | $0.00016 | $0.03367 |
| Sonnet 5 | $0.00006 | $0.01347 |
| Haiku 4.5 | $0.00003 | $0.00673 |
Grade A, and why
unity-performance 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 725 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Unity Performance
Deep decision framework for Unity performance analysis and optimization. Profile first, identify the bottleneck, then apply the targeted fix. This skill provides CPU/GPU/RAM cost models, object pooling, batching, physics, and platform budget guidance.
When to Apply
Activate when the conversation involves:
- Profiling or measuring performance in the Unity Editor or on-device
- Choosing between rendering strategies (static batching, GPU instancing, SRP Batcher, atlases)
- Deciding whether to pool objects, and how to implement the pool
- Physics performance (collision layers, collider types, timestep tuning)
- Memory management (texture compression, audio load types, heap allocations)
- Frame rate targets and platform budget planning
- Identifying hotspots in Update loops or high-frequency code paths
Do NOT activate for broad architectural questions about how to structure a system — those go to hades:unity-architect first.
Project Context Check
Before making recommendations, gather project-specific context so advice is calibrated to the actual codebase.
-
Understand the project scale and render pipeline:
- Call
get_project_summary()— reveals render pipeline (URP/HDRP/Built-in), platform target, scene count, and asset volumes. A mobile URP project requires very different advice than a PC HDRP project. - Call
project_settings(section: "renderPipeline")— confirms which pipeline is active and its pipeline asset path. Hades has no tool that reports SRP Batcher status, active rendering features, or custom passes — check those directly in Project Settings > Graphics or the Frame Debugger.
- Call
-
Find hot-path candidates:
- Call
graph_query(edgeKind: "references", edgeTargetNamePattern: "Update", edgeTargetKind: "Class")— lists components referencing a script whose name matches "Update". Every entry is a hot-path candidate. Large counts on mobile are an immediate concern. - Call
graph_query(edgeKind: "references", edgeTargetNamePattern: "FixedUpdate", edgeTargetKind: "Class")— physics-side hot paths. - Call
graph_query(edgeKind: "references", edgeTargetNamePattern: "LateUpdate", edgeTargetKind: "Class")— camera/IK hot paths.
- Call
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.
- 2d ago First seen · 725 lines · 32 tokens per session scan A 53acd6d35714
unity-performance is a skill published in the GitHub repository TheArcForge/Hades (27 stars, last pushed 6d ago), licensed MIT. It adds 32 tokens to every session and 6,734 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
unity-reviewer
Use after implementation to review code changes for Unity anti-patterns, performance issues, and correctness problems. Also use when superpowers:requesting-code-review is active to provide Unity-specific review criteria.
game-developer
Use when building game systems, implementing Unity/Unreal features, or optimizing game performance. Invoke for Unity, Unreal, game patterns, ECS, physics, networking, performance optimization.
component-design
Use when you need detailed decision frameworks for MonoBehaviour composition, inter-component communication, single responsibility, and execution order.
data-modeling
Use when you need detailed decision frameworks for ScriptableObject patterns, runtime data management, serialization strategies, and data-driven architecture.
prefab-architecture
Use when you need detailed decision frameworks for prefab vs variant vs ScriptableObject, nested prefab design, override strategies, and prefab editing workflows.
scene-architecture
Use when you need detailed decision frameworks for scene structure, additive loading, scene management, transitions, and persistent managers.