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.
git clone --depth 1 https://github.com/j4flmao/agent-skillsWrote 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/rules/j4flmao/agent-skills/game-engine-guidelines)<a href="https://agentmods.dev/rules/j4flmao/agent-skills/game-engine-guidelines"><img src="https://agentmods.dev/badge/rules/j4flmao/agent-skills/game-engine-guidelines/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/j4flmao/agent-skills/game-engine-guidelines"><img src="https://agentmods.dev/badge/rules/j4flmao/agent-skills/game-engine-guidelines.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00020 | $0.01456 |
| Opus 5.5 | $0.00008 | $0.00582 |
| Sonnet 5.5 | $0.00004 | $0.00291 |
| Haiku 4.5 | $0.00002 | $0.00146 |
Grade A, and why
game-engine-guidelines 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 1mo 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Game Engine Guidelines
AI/Cursor MUST follow these rules when writing game engine, renderer, or gameplay systems code. These rules harden correctness, performance, and determinism.
1. Performance & Frame Budget
- Rule: Profile before optimizing; never guess. Every hot path must include measurable metrics (
TimeMeasure, GPU timers, counters). - Rule: Allocators in the render/hot loop are forbidden. Use
pool,stack, orarenaallocators per frame; reserve at startup. - Rule: Draw calls are precious. Batch by material/pipeline; cap state changes (
SetPipeline/SetShader); prefer instancing and indirect draw. - Rule: No
new/mallocduring simulation or render ticks. If code allocates in a tick, it is a review-blocker. - Rule: Vectors of
unique_ptrto polymorphic entities are rejected. Keep component data in contiguous arrays (SoA) and entity IDs index them.
2. Data-Oriented & Cache Design
- Rule: Model CPU cache: elements in hot systems are contiguous, 32-64 byte stride structs; split hot/cold component data into separate arrays.
- Rule: Prefer
std::vector<T>+ index handles over object pointers in entity systems. - Rule: Object-oriented hierarchies (
Actorbase class with virtualTick()) are allowed only for gameplay-throttled, low-frequency logic; never per-entity per-frame. - Rule: SIMD is required where a uniform operation runs over many elements (transforms, particle sim). Write scalar reference first, then vectorize.
3. Threading & Determinism
- Rule: The simulation step MUST be deterministic given the same inputs (same seed, same binary, same settings). Never allow wall-clock time, map iteration order, or
std::rand()into the simulation. - Rule: Render jobs, physics, and sim run on a job graph; no thread may block on another for more than a bounded slice.
- Rule: All shared mutable state is owned by exactly one system (ownership), or accessed via
atomic/lock-free structures that are benchmarked. Unlocked cross-thread writes = review-blocker. - Rule:
Tickdata is never mutated while being read by another thread; use per-frame double buffering or generation counters. - Rule: Never block the render thread on asset I/O, shader compile, network, or GC.
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.
- 1mo ago First seen · 88 lines · 20 tokens per session scan A 48813ad4d694
game-engine-guidelines is a cursor rule published in the GitHub repository j4flmao/agent-skills (33 stars, last pushed 10d ago), licensed MIT. It adds 20 tokens to every session and 1,456 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-09-20.
Other cursor rules, from other repositories
meta-quest-agentic-tools
Use Meta Quest Agentic Tools for Meta Quest and Horizon OS samples.
narrative-writing
A set of rules for writing and editing video-game narratives, connected to a narrative project. It requires the agent to load the project's writing context and story canon, which is the established set of story facts, before making changes.
meta-quest-agentic-tools
Use Meta Quest Agentic Tools for Meta Quest and Horizon OS samples.
visual-and-observational-rules
Defines the visual aspects of the game and how the player observes the world. This includes map color-coding, screen effects, and the overall simulation style.
00-project-overview
Base project rule — purpose, constraints, and defaults for mcp-5etools.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.