playbooks

home / skills / a5c-ai / babysitter / destruction

destruction skill

/plugins/babysitter/skills/babysit/process/specializations/game-development/skills/destruction

This skill enables dynamic scene destruction by fracturing meshes, simulating debris and propagation while managing performance for realistic environments.

npx playbooks add skill a5c-ai/babysitter --skill destruction

Review the files below or copy the command above to add this skill to your agents.

Files (2)
SKILL.md
326 B
---
name: destruction
description: Destructible environment skill for fracturing and debris systems.
allowed-tools: Read, Grep, Write, Bash, Edit, Glob, WebFetch
---

# Destruction Skill

Destructible environments and objects.

## Capabilities

- Mesh fracturing
- Debris physics
- Damage propagation
- Performance management

Overview

This skill provides a JavaScript-ready destructible environment system focused on mesh fracturing, debris simulation, and controlled damage propagation. It integrates performance-aware features so fracturing and debris remain deterministic and resumable within orchestrated workflows. Use it to add believable breakage and rubble behavior to interactive scenes, simulators, or automated testing of physics-driven systems.

How this skill works

The skill fractures meshes into pieces according to configured patterns and material properties, then spawns debris bodies with physical properties and constraints. Damage inputs propagate through connected pieces using configurable decay and threshold rules, producing realistic chain reactions while keeping CPU and memory usage bounded. Deterministic seeding and resumable state capture make the system suitable for automated orchestration and replay in development pipelines.

When to use it

  • Adding realistic breakage and rubble to games, VR, or AR scenes.
  • Simulating structural failure for training, testing, or validation.
  • Automating physics-based regression tests where repeatability is required.
  • Creating cinematic destruction sequences with performance constraints.
  • Prototyping material responses and damage propagation in research.

Best practices

  • Precompute fracture patterns and LOD variants to reduce runtime cost.
  • Limit active debris count and use pooling to avoid GC spikes.
  • Tune damage thresholds and propagation decay to match physical scale.
  • Use deterministic seeds and snapshot state for resumable orchestration.
  • Profile on target hardware and enable progressive fracture/detail levels.

Example use cases

  • A game boss fight where environment pieces shatter and cascade realistically when hit.
  • An AR demo that fractures props on impact while keeping framerate on mobile.
  • Automated test that applies repeatable damage sequences to verify repair algorithms.
  • A cinematic tool that exports deterministic destruction timelines for rendering passes.

FAQ

Is the destruction deterministic across runs?

Yes — the system supports deterministic seeds and state snapshots so fracture results and debris behavior can be replayed exactly.

How does it keep performance predictable?

It uses LOD, debris pooling, capped active fragment counts, and progressive fracture so you can trade fidelity for steady CPU and memory usage.