unrealxu/unrealengine5-skills/ue5-performance-packaging

Claude Code and agent skill by unrealxu/unrealengine5-skills

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

UE5.6-UE5.8 performance and packaging readiness workflow. Use when requests involve PIE performance checks, runtime stat review, pre-package validation, build configuration sanity, and release readiness checklists.

Install

npx vskill@latest install unrealxu/unrealengine5-skills/ue5-performance-packaging

SKILL.md

Quick Start

  • Confirm target platform and build configuration.
  • Collect current performance symptoms and packaging goal.
  • Output a pre-package checklist plus measurement plan.

API Anchors (UE5.6-UE5.8)

  • Runtime quality and frame-budget anchors:
  • UGameUserSettings::SetOverallScalabilityLevel(...)
  • UGameUserSettings::SetFrameRateLimit(...)
  • UGameUserSettings::ApplySettings(...)
  • Packaging settings anchors:
  • UProjectPackagingSettings
  • BuildConfiguration, MapsToCook, DirectoriesToAlwaysCook
  • UsePakFile, bUseIoStore, bGenerateChunks
  • Asset dependency validation anchors:
  • FAssetRegistryModule, IAssetRegistry
  • GetAssetsByPath(...)
  • GetDependencies(...), GetReferencers(...)
  • Profiling instrumentation anchors:
  • TRACE_CPUPROFILER_EVENT_SCOPE(...)
  • DECLARE_CYCLE_STAT(...)
  • Practical runtime stat commands:
  • stat unit, stat gpu, stat scenerendering, memreport -full

Performance and Packaging Contract

  • Every performance/packaging task must define:
  • target platform + build config + test map
  • reproducible capture scenario (camera path, duration, net mode)
  • baseline metrics and acceptance thresholds
  • packaging configuration set and dependency scan scope
  • go/no-go output with explicit blockers
  • If any item is missing, readiness evaluation is incomplete.

Workflow

1) Reproducible Baseline Capture

  • Freeze test map, camera path, scalability, and net mode.
  • Capture stat unit and stat gpu under same scenario repeatedly.
  • Use median or percentile metrics, not single-frame spikes.

2) Hotspot Isolation

  • Identify top-cost gameplay/render systems in the capture window.
  • Correlate high-cost actors/assets with scene context.
  • Add scoped CPU markers for custom systems when attribution is unclear.

3) Asset and Dependency Validation

  • Scan target paths for missing or broken asset references.
  • Query dependencies and referencers for problematic assets.
  • Validate required map/game mode assets are included in package scope.

4) Packaging Configuration Validation

  • Check UProjectPackagingSettings fields for target release policy.
  • Validate maps-to-cook, always-cook directories, and build configuration.
  • Validate Pak/IoStore/chunk strategy against distribution requirements.

5) Package Readiness Trial

  • Run pre-package checklist and dry-run style verification.
  • Isolate first blocking packaging error and dependency chain.
  • Record unresolved blockers with owner and risk.

6) Go/No-Go Decision

  • Approve only when metrics and packaging checks meet thresholds.
  • Report remaining risks and recommended mitigation actions.
  • Lock scenario and settings used for sign-off evidence.

Constraints

  • Do not claim optimization wins without measurable before/after data.
  • Keep profiling scenario reproducible (map, camera path, net mode).
  • Distinguish editor overhead from packaged runtime behavior.
  • Avoid changing quality scalability settings without reporting it.
  • Keep packaging checks deterministic; avoid ad-hoc config toggles during sign-off.
  • Always report metric units (ms, fps, memory) and capture duration.

Failure Handling

  • Symptom: performance data is noisy across runs.
  • Locate: non-deterministic scenario setup and background variance.
  • Fix: lock scenario, warm up run, and compare median/percentile values.
  • Symptom: optimization claims do not reproduce.
  • Locate: mismatched quality settings, map, or runtime mode.
  • Fix: publish exact capture config and rerun before/after under same conditions.
  • Symptom: package build fails with long error cascade.
  • Locate: first blocking error and its direct dependency chain.
  • Fix: resolve earliest blocker first; rerun to reveal next blockers.
  • Symptom: packaged build launches with missing assets.
  • Locate: cook list coverage and dependency graph gaps.
  • Fix: include missing maps/directories and resolve broken references.
  • Symptom: packaged build perf is worse than PIE baseline.
  • Locate: packaged-only config differences and runtime content path.
  • Fix: compare packaged config to baseline and align scalability/device settings.
  • Symptom: memory usage regresses near content-heavy scenes.
  • Locate: high-memory assets and streaming policy behavior.
  • Fix: reduce resident asset pressure and adjust streaming/cook strategy.

Packaging Ops

  • Prefer explicit map/cook lists over implicit discovery for release builds.
  • Use AssetRegistry queries to validate dependencies before packaging.
  • Keep one source of truth for release packaging settings per target profile.
  • In UE5.8, treat Zenserver cooked output as an iteration store; keep Pak/IoStore staging validation for distributable builds.
  • Treat UE5.8 Incremental Cooking as a beta iteration path and retain a clean/full-cook release check.
  • Re-run readiness checks after any packaging setting change.

UE5.6-UE5.8 Compatibility Notes

  • UGameUserSettings, UProjectPackagingSettings, and AssetRegistry APIs above are stable in UE5.6-UE5.8.
  • ProjectPackagingSettings.h lives under Developer/DeveloperToolSettings across UE5.6-UE5.8.
  • UE5.8 enables Zenserver as the cooked output store by default for supported iteration workflows; this does not replace release container validation.

Escalation

  • Escalate when performance bottleneck requires engine-level profiling or renderer changes.
  • Escalate when packaging issues stem from third-party plugin build failures.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
278
Repository forks
38
7-day trend score
28

Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 21 Mar 2026, 7 months ago.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/unrealxu/unrealengine5-skills/ue5-performance-packaging/badge)](https://verified-skill.com/skills/unrealxu/unrealengine5-skills/ue5-performance-packaging)

Something wrong with this skill or its page? Tell us.