unreal-build-vfx

Author and integrate Unreal Niagara effects with bounded simulation, explicit pooling, cancellation and target scalability evidence.

SKILL.md

GitHub

Unreal Build VFX

When to use

Build or integrate Unreal Niagara effects, their runtime parameter adapters, pooling, bounds and scalability. Shared gameplay feedback design remains agnostic; this feature owns Niagara assets and native runtime integration.

Activation and prerequisites

Run commands/resolve-engine.sh and commands/resolve-skill.sh unreal-build-vfx before writes or tool launches. Require effective engine UNREAL and compatible: true, with validated override/output scope when present. Skill selection never selects an engine.

Inspect engine version, selected Niagara plugin/modules, CPU/GPU simulation, renderer/platform support, required system assets, event sources, quality levels and effect budgets. Do not enable plugins, choose cloud media or install dependencies implicitly.

Workflow

  1. Specify each effect's semantic event, authority/prediction policy, owner, transform/attachment and completion/cancellation route. Deduplicate predicted/confirmed presentation without letting visual timing determine hits.
  2. Author genuine Niagara systems/emitters, materials and parameter interfaces with native tools. Preserve source provenance and art direction; inspect parameter names/types against real assets.
  3. Select CPU/GPU simulation, bounds, spawn/update rates, collision and data interfaces against target support. Declare world versus local space and explicit game/real/sequence clock behavior.
  4. Choose component or system-spawn ownership and pool method deliberately. Configure parameters and bindings before activation, then retire handles safely; use the Niagara runtime reference for API and pool lifecycle decisions.
  5. Stop loops and trails on cancellation, replacement, actor teardown or travel. Reset every owned parameter, attachment and delegate on reuse; never mutate a component after releasing it to the pool.
  6. Use NiagaraEffectType/scalability and measured instance/particle/overdraw limits. Define how culling affects purely visual events and keep required gameplay information available through another channel.

Verification

Test overlapping events, pool reacquisition, stale completion, actor removal, travel and all selected clocks. Inspect near/far views, bounds, pause, low frame rates and scalability tiers in native captures. Measure CPU/GPU, draw/overdraw and peak instance counts on the actual target. A finite CPU-emitter pool test does not certify GPU effects or appearance.

Output

Native effect assets, typed parameter and event contracts, pool/reset/cancel rules, target budgets, lifecycle tests and native visual/performance findings. Record environment, expected/observed results and blocked/not-run checks under the active .hephaera/.cache/tasks/ task.

Boundaries

Preserve media, provenance, independent review and promotion controls. Headless success cannot certify effect readability, flashing safety, appearance or physical haptics.

References

  • Niagara runtime: typed parameters, pooling, completion, scalability and acceptance.