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/teddy563/mcwrench/performance-tuningnpx skills add Teddy563/mcwrench --skill performance-tuninggit clone --depth 1 https://github.com/Teddy563/mcwrenchWrote 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/teddy563/mcwrench/performance-tuning)<a href="https://agentmods.dev/skills/teddy563/mcwrench/performance-tuning"><img src="https://agentmods.dev/badge/skills/teddy563/mcwrench/performance-tuning.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 | $0.00191 | $0.00934 |
| Opus 5 | $0.00096 | $0.00467 |
| Sonnet 5 | $0.00038 | $0.00187 |
| Haiku 4.5 | $0.00019 | $0.00093 |
Grade A, and why
performance-tuning 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 4d 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Performance Tuning
Diagnose first, tune second. Never hand someone a wall of flags without measuring.
Step 1 — Measure with Spark
Spark (spark.lucko.me) is the modern profiler (replaces Timings; Paper bundles it). See
references/spark-cheatsheet.md. The minimum diagnostic loop:
/spark tpsand/spark health— is it a TPS problem, an MSPT spike problem, or memory?/spark profiler start --timeout 300during the lag, then read the uploaded flame graph.- For memory pressure / GC churn:
/spark profiler start --alloc,/spark gc,/spark heapsummary.
Identify whether the cost is the main thread (ticking: entities, hoppers, redstone, chunk gen), GC, or I/O (world saves, plugin storage on slow disk).
Step 2 — Heap & JVM flags
See references/aikars-flags.md for the full verified flag set and sizing rules. Key points:
Xms == Xmx. Leave 1–1.5 GB headroom for the OS + JVM native memory beyond the heap.- Aikar's flags are G1GC-tuned and remain the proven default on the official PaperMC page.
- Java 25 nuance: Paper 26.1+ requires Java 25. Hosting/community guidance (e.g. WinterNode — NOT PaperMC's own docs) recommends not pairing Aikar's G1 flags with Java 25's Generational ZGC; choose one GC. Present Aikar's/G1 as the safe default and ZGC as an alternative to benchmark, and never mix the two flag sets.
- On memory-capped containers (Pterodactyl/Pelican),
-XX:+AlwaysPreTouchcan cause "Cannot allocate memory" at boot — drop it there.
Step 3 — Config tuning (biggest wins first)
In observed frequency order:
view-distance/simulation-distance(server.properties,spigot.yml, per-worldpaper-world.yml). The lower of the configured values wins per world. Typical sane values: view 6–10, simulation 4–6. Anarchy/huge servers go lower.- Mob farms / entity load —
spigot.yml: entity-activation-range,mob-spawn-range, Paper per-player mob spawns; cap withentity-per-chunk-save-limitforexperience_orb,arrow,ender_pearl,item. - Hoppers —
paper-world.ymlhopper tuning /transfer-cooldown; reduce move-event overhead on hopper-heavy economies. - Chunk storms — cap
player-max-chunk-load-rate(~100.0) for exploration-heavy servers. - Redstone clocks / lag machines — find them in the Spark profile and address the build.
- Sync I/O — move LuckPerms/world storage off slow disks; stagger autosaves.
What ships with it
2 files 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.
- 4d ago First seen · 67 lines · 191 tokens per session scan A 28f81e136f20
performance-tuning is a skill published in the GitHub repository Teddy563/mcwrench (1 stars, last pushed 11d ago), licensed MIT. It adds 191 tokens to every session and 934 once invoked, about $0.0010 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 2026-08-31.
Other skills, from other repositories
develop-web-game
Use when Codex is building or iterating on a web game (HTML/JS) and needs a reliable development + testing loop: implement small changes, run a Playwright-based test script with short input bursts and intentional pauses, inspect screenshots/text, and review console errors with rendergametotext.
ai-behavior-trees-utility-ai
Build a production behavior-tree runtime (Blackboard, action/condition leaves, sequence/selector/parallel composites, decorators) and a Utility AI system (response curves — linear, exponential, sigmoid, quadratic — considerations, and action evaluators), plus hybrid BT-drives-Utility agents. Use when implementing a…
camera-systems
Build game cameras that feel good — 2D follow with a deadzone, look-ahead, smoothing, and level-bounds clamping; 3D third-person orbit with collision and first-person look; plus multi-target framing and a shake hook. Engine-neutral techniques that pair with the engine's camera node and rigs like Unity Cinemachine or…
create-game-assets
Plan, generate, source, normalize, and validate cohesive visual game assets. Use for art direction, style bibles, sprites, tilesets, backgrounds, UI art, icons, textures, concept art, or 3D asset briefs.
game-ai
Design NPC and enemy decision-making with finite state machines, behavior trees, steering behaviors, and A pathfinding — engine-neutral algorithms that pair with the detected engine's navigation API. Use when building enemy AI, an FSM or behavior tree, steering/flocking, or pathfinding, or when the user mentions state…
game-feel
Add "juice" and game feel that makes actions satisfying — screen shake, hit-stop/freeze frames, tweened/eased motion, squash & stretch, knockback, and layered audio-visual feedback — as engine-neutral techniques that pair with the detected engine's tween, particle, and camera APIs. Use when the user mentions game…