
unreal-behavior-trees
Build NPC AI in Unreal Engine 5 with Behavior Trees and Blackboards: composites (Selector/Sequence), tasks, decorators, services, and running the tree from an AIController. Use when creating enemy/NPC AI, BT/BB assets, custom BTTask or BTService nodes, or when the user mentions Behavior Tree, Blackboard, AIController, BTTask, decorator, or service.
- Installs
- 5.5K
- Last 24h
- +294
- Trending rank
- #978
- Security
- Pass
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/gamedev-skills/awesome-gamedev-agent-skills --skill unreal-behavior-trees- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads unreal-behavior-trees automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
Security audits
Independent scans of the skill's instructions and bundled files.
SKILL.md
The instructions your agent reads when the skill is triggered.
Unreal Behavior Trees
Author NPC decision-making in UE5 with Behavior Trees driven by a Blackboard: structure the tree with composites, gate branches with decorators, keep state current with services, and run it from an AIController. Targets UE 5.8.
When to use
- Use when building enemy/NPC AI: creating a
BT_/BB_asset pair, structuring Selector/Sequence branches, adding decorators (conditions) and services (periodic updates), writing customBTTask/BTServicenodes, or wiring an AIController to run the tree. - Use when the project has Behavior Tree (
BT_) and Blackboard (BB_) assets and anAAIController.
When not to use: the concept of AI (FSM vs BT vs steering, cross-engine) →
game-ai. Pure navigation/pathing math is engine navmesh (BT's MoveTo uses it). Simple
one-off logic may be cheaper as a small state machine than a full tree.
Behavior Tree or StateTree? UE 5.8 ships both, and Behavior Trees are not deprecated.
StateTree is Epic's general-purpose hierarchical state machine: it combines BT-style
selection with explicit States and Transitions. If the project already uses StateTree
assets, or the AI is naturally a set of modes with clear transitions (patrol → alert →
combat), check with the user before adding a Behavior Tree next to it. Keep using Behavior
Trees where the project already has BT_/BB_ assets.
Core workflow
- Create the pair: a Blackboard (
BB_) holds typed keys (the AI's memory:TargetActor,LastKnownLocation,bIsInvestigating); a Behavior Tree (BT_) references that Blackboard. - Possess and run. An
AAIControllerpossesses the pawn and callsRunBehaviorTree(BT), which also initializes the referenced Blackboard. - Structure with composites. Selector runs children left→right until one succeeds (priority/fallback: "attack, else chase, else patrol"). Sequence runs children until one fails (do-all: "move to cover → reload → peek"). Simple Parallel runs one main task alongside a secondary.
- Gate branches with Decorators that read Blackboard keys (e.g. "Has Target?" guards the combat branch). Set Observer Aborts so the tree re-evaluates when the key changes.
- Keep the Blackboard current with Services attached to a branch — they tick periodically
(e.g. update
TargetActorvia a sight check) only while that branch is active. - Do work in Tasks, which return
Succeeded,Failed, orInProgress(latent tasks likeMoveTofinish later). - Verify with the Behavior Tree debugger during PIE — it highlights the running node and shows live Blackboard values, so you see exactly which branch executes.
Patterns
1. AIController that runs the tree (C++)
void AEnemyAIController::OnPossess(APawn* InPawn)
{
Super::OnPossess(InPawn);
if (BehaviorTree) // UPROPERTY(EditAnywhere) TObjectPtr<UBehaviorTree>
RunBehaviorTree(BehaviorTree); // initializes & uses the Blackboard the BT references
}
2. A priority tree (node structure)
ROOT
└── Selector (try combat, else investigate, else patrol)
├── Sequence [Decorator: Blackboard 'TargetActor' Is Set, Observer Aborts: Both]
│ ├── Task: MoveTo (TargetActor) // latent: returns InProgress then Succeeded
│ └── Task: Attack
├── Sequence [Decorator: 'LastKnownLocation' Is Set]
│ ├── Task: MoveTo (LastKnownLocation)
│ └── Task: Wait (3s) + clear key
└── Task: Patrol (BTTask_FindPatrolPoint -> MoveTo)
Observer Aborts: Both makes the combat branch interrupt patrol the instant TargetActor is
set, and bail out when it's cleared — this is what makes the AI feel reactive.
3. Updating the Blackboard from code (e.g. on seeing the player)
void AEnemyAIController::SetTarget(AActor* Target)
{
if (UBlackboardComponent* BB = GetBlackboardComponent())
BB->SetValueAsObject(TEXT("TargetActor"), Target); // key name must match the BB asset
}
// Clear with BB->ClearValue(TEXT("TargetActor")); to drop back to a lower-priority branch.
Pitfalls
- AI never starts — the pawn isn't possessed (set the Pawn's Auto Possess AI to "Placed
in World or Spawned" and assign the AIController), or
RunBehaviorTreewas never called. MoveToinstantly fails — no NavMesh in the level (add a Nav Mesh Bounds Volume), or the target is off the navmesh.- Branch doesn't react to changes — the gating Decorator's Observer Aborts is set to None; set it to Self/Lower Priority/Both so the tree re-evaluates when the key changes.
- A task hangs the tree — a custom task returned
InProgressand never callsFinishLatentTask. Always complete latent tasks. - AI behaves strangely only when several NPCs share a tree — a custom task stores run state in member variables. Task nodes are shared by every AI using the asset unless instanced; keep per-AI state in node memory (see the reference).
- Blackboard key typos —
SetValueAsObject("Taget", ...)silently does nothing; match the key name and type exactly, or use a cachedFBlackboardKeySelector. - Sequence vs Selector confusion — Sequence = AND (stops on first failure); Selector = OR (stops on first success). Swapping them inverts the behaviour.
References
- For a custom C++
UBTTaskNode(instant and latentExecuteTaskreturningEBTNodeResult, per-AI node memory,AbortTask, and aFBlackboardKeySelector), readreferences/custom-bttask.md. - Primary docs: "Behavior Trees in Unreal Engine"
(
https://dev.epicgames.com/documentation/en-us/unreal-engine/behavior-trees-in-unreal-engine).
Related skills
game-ai— engine-agnostic AI design (FSM, BT, steering, pathfinding choices).unreal-cpp-gameplay— the AIController and pawn classes in C++.fps-shooter/tower-defense— genres that compose enemy AI.
FAQ
What does the unreal-behavior-trees skill do?
Build NPC AI in Unreal Engine 5 with Behavior Trees and Blackboards: composites (Selector/Sequence), tasks, decorators, services, and running the tree from an AIController. Use when creating enemy/NPC AI, BT/BB assets, custom BTTask or BTService nodes, or when the user mentions Behavior Tree, Blackboard, AIController, BTTask, decorator, or service.
How do I install unreal-behavior-trees?
Run npx skills add https://github.com/gamedev-skills/awesome-gamedev-agent-skills --skill unreal-behavior-trees in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/unreal-behavior-trees/ (personal) or .claude/skills/unreal-behavior-trees/ (project).
Which AI agents can use unreal-behavior-trees?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is unreal-behavior-trees safe to install?
Security partners have reviewed it: Gen Agent Trust Hub passed; Socket passed; Snyk passed. As with any skill, read the SKILL.md and bundled scripts before installing.
Who maintains unreal-behavior-trees?
It is published by gamedev-skills in gamedev-skills/awesome-gamedev-agent-skills. It has 5,454 recorded installs.
More from gamedev-skills/awesome-gamedev-agent-skills
game-feel
gamedev-skills/awesome-gamedev-agent-skills
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 feel, juice, "make it feel good/punchy", screen shake, hit stop, screen freeze, easing, squash and stretch, impact frames, or feedback/polish on hits, jumps, pickups, and deaths.
game-ui-ux
gamedev-skills/awesome-gamedev-agent-skills
Design and build game UI/UX — HUDs, menus, and overlays — that survive every screen: anchor- based responsive layout, resolution/aspect scaling and safe areas, keyboard/gamepad focus navigation, a screen/menu state stack, and event-driven (not polled) HUD updates. Engine- neutral patterns that pair with the detected engine's UI skill. Use when the user mentions HUD, health bar, main menu, pause menu, settings screen, UI layout, anchors, UI scaling, aspect ratio, safe area, controller/keyboard menu navigation, or wiring UI to game state.
level-design
gamedev-skills/awesome-gamedev-agent-skills
Design and build playable levels — the blockout/whitebox-to-playable workflow, player metrics and grid layout, pacing and flow (tension/rest curve), gating and the critical path, and encounter design. Engine-neutral practice. Use when the user mentions level design, blockout/whitebox/greybox, level layout, level pacing, encounter design, or the critical path through a level.
performance-optimization
gamedev-skills/awesome-gamedev-agent-skills
Find and fix game performance problems methodically — measure with the engine profiler first, reason about the frame-time budget, locate the CPU-vs-GPU bottleneck, then apply the right fix: object pooling, draw-call batching, fewer allocations/GC spikes, and asset budgets. Engine- neutral method that pairs with each engine's profiler. Use when the user mentions performance, optimize, low/dropping FPS, frame drops, stutter, lag, profiler, frame budget, draw calls, batching, garbage collection/GC spikes, object pooling, or "the game runs slow".
game-ai
gamedev-skills/awesome-gamedev-agent-skills
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 machine, behavior tree, blackboard, A, navmesh, seek, or patrol/chase.
audio-design
gamedev-skills/awesome-gamedev-agent-skills
Implement game audio practice — bus/mixer architecture and gain in decibels, ducking (sidechain), adaptive/dynamic music via layering and re-sequencing, SFX variation, and beat synchronization. Engine-neutral. Use when the user mentions audio mixing, audio buses, adaptive/dynamic music, ducking, SFX variation, music layers, or syncing gameplay to the beat.
Popular ai & agents skills
google-agents-cli-publish
google/agents-cli
This skill should be used when the user wants to "publish an agent", "publish my ADK agent", "register an agent with Gemini", "publish to Gemini" (formerly Gemini Enterprise), or needs guidance on the agents-cli publish gemini command. Also use when the user wants to "manage agents in Agent Registry", "list/update/delete registered agents", or "register an MCP server". Covers ADK vs A2A registration modes, programmatic and interactive usage, flag reference, auto-detection from deployment metadata, Agent Registry fleet management, and troubleshooting. Part of the agents-cli skills suite. Do NOT use for deployment (use google-agents-cli-deploy).
google-agents-cli-eval
google/agents-cli
This skill should be used when the user wants to "run an evaluation", "evaluate my agent", "evaluate my ADK agent", "write an eval dataset", "analyze eval failures", "compare eval results", "optimize agent", or needs guidance on the Gemini platform eval methodology and the Quality Flywheel. Covers eval metrics, dataset schema, LLM-as-judge scoring, and common failure causes. Applies to any agents-cli project, whatever framework the agent is written in. Do NOT use for agent API code patterns (ADK: use google-agents-cli-adk-code), deployment (use google-agents-cli-deploy), or project scaffolding (use google-agents-cli-scaffold).
skill-creator
anthropics/skills
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
codex-pet
prime-skills/runcomfy-agent-skills
Codex Pet generator on RunComfy. Build a Codex-compatible Codex Pet spritesheet.webp + pet.json from a single reference image, drop it into ${CODEXHOME:-$HOME/.codex}/pets/ / and Codex picks it up as a custom Codex Pet next to the 8 built-ins. This skill produces the exact Codex Pet atlas Codex expects (1536x1872 PNG/WebP, 8 cols x 9 rows, 192x208 cells, 9 animation states — idle, running-right, running-left, waving, jumping, failed, waiting, running, review). Calls OpenAI GPT Image 2 edit ONCE via the local RunComfy CLI as runcomfy run openai/gpt-image-2/edit to produce a canonical Codex Pet pose, then assembles all 9 animation rows programmatically with ImageMagick micro-transforms — no Codex Pro, no $imagegen, no OPENAIAPIKEY required, only RUNCOMFYTOKEN. Triggers on "codex pet", "create codex pet", "make codex pet", "hatch codex pet", "/hatch image", "desktop pet codex", "codex pets", "spritesheet.webp", or any explicit ask to build a custom pet for OpenAI Codex.
full-output-enforcement
leonxlnx/taste-skill
Overrides default LLM truncation behavior. Enforces complete code generation, bans placeholder patterns, and handles token-limit splits cleanly. Apply to any task requiring exhaustive, unabridged output.
executing-plans
obra/superpowers
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available