Godot Genre Fighting logo

Godot Genre Fighting

Community
thedivergentai
godot-genre-fighting

Expert blueprint for fighting games including frame data (startup/active/recovery frames, advantage on hit/block), hitbox/hurtbox systems, input buffering (5-10 frames), motion input detection (QCF, DP), combo systems (damage scaling, cancel hierarchy), character states (idle/attacking/hitstun/blockstun), and rollback netcode. Based on FGC competitive design. Trigger keywords: fighting_game, frame_data, hitbox_hurtbox, input_buffer, motion_inputs, combo_system, rollback_netcode, cancel_system, advantage_frames.

Overview

Publisherthedivergentai
RepositoryGD-Agentic-Skills
Skill namegodot-genre-fighting
Stars
727
Forks
43
Bundled files
21
LicenseLGPL-3.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 21 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by thedivergentai on GitHub. Read the source before you install it.

Installation

Install the Godot Genre Fighting AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/thedivergentai/GD-Agentic-Skills.git /tmp/GD-Agentic-Skills
mkdir -p .claude/skills
cp -r /tmp/GD-Agentic-Skills/skills/godot-genre-fighting .claude/skills/godot-genre-fighting
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Godot Genre Fighting in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Godot Genre Fighting on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Godot Genre Fighting is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

NEVER Do (Expert Anti-Patterns)

Frame-Data & Logic

  • NEVER use variable framerates; strictly lock logic to a Deterministic Fixed Loop (using _physics_process with a frame-counter) and call reset_physics_interpolation() on teleport.
  • NEVER use standard Physics for hit detection; strictly use PhysicsDirectSpaceState.intersect_shape() to query hitboxes instantly without Area2D signal lag.
  • NEVER skip Damage Scaling; strictly apply 10% reduction per hit in a combo to prevent infinite matches.
  • NEVER make all moves safe on block; strictly ensure high-reward moves have Recovery Windows where the attacker is punishable.
  • NEVER rely on Area2D.get_overlapping_areas(); strictly use intersect_shape() for immediate, frame-perfect resolution.
  • NEVER forget Hitbox Proximity (Proximity Guard); strictly trigger guard states when a hitbox enters a nearby zone, even if it hasn't landed.

Character & Animation

  • NEVER use simple parenting (scale.x = -1) for character flip; strictly adjust the dedicated Visuals node while managing hitbox offsets programmatically.
  • NEVER use string-based animation triggers; strictly use AnimationMixer with ADVANCE_MANUAL for frame-synced playback.
  • NEVER use yield or await for frame-critical logic; strictly use Integer Frame Counting within state machines to manage recovery/startup windows perfectly.
  • NEVER store frame data in raw scripts; strictly use Resource files (.tres) with delegated logic for damage scaling, cancels, and combo-state tracking.
  • NEVER use deep node hierarchies for character parts; strictly keep skeletons shallow to reduce transformation overhead.

Input & Networking

  • NEVER skip Input Buffering; strictly implement a 5-10 frame buffer to ensure lenient, responsive execution for the player.
  • NEVER leave Input.use_accumulated_input enabled; strictly disable it to preserve sub-frame timing for precise combo links.
  • NEVER use client-side hit detection for netplay; strictly use rollback netcode or server validation to prevent desyncs.
  • NEVER use standard TCP for multiplayer; strictly use UDP/ENet to avoid head-of-line blocking during latency spikes.
  • NEVER rely on the SceneTree for fighter transforms in netplay; strictly manage positions in a serializable data buffer.

🛠 Expert Components (scripts/)

MANDATORY reads before implementing the matching system:

  1. fighting_input_buffer.gd — buffers + motion (QCF/DP)
  2. direct_hitbox_query.gdexclusive hit resolution path (intersect_shape)
  3. rollback_state_serializer.gd — snapshot/restore for netplay

Original Expert Patterns

  • fighting_input_buffer.gd - Frame-locked input engine (60fps) with motion command fuzzy matching (QCF/DP).
  • hitbox_component.gd - Hitbox/hurtbox volume helper (layers High/Low/Throw) — resolve hits via direct_hitbox_query, not Area signals.

Modular Components

Restored from baseline (load on demand)


Core Loop

Neutral → Confirm Hit → Combo → Advantage → Repeat

Decision Trees (no Area2D / inline system dumps)

Frames & fixed loop

NeedAction
Attack timing Resource.tres with startup/active/recovery/advantage — not script constants
60fps sim stepdeterministic_physics_loop.gd
Anim sync to framesmanual_animation_advancer.gd (ADVANCE_MANUAL)

Input

NeedAction
5–10f buffer + QCF/DPMANDATORY fighting_input_buffer.gd
Sub-frame linksinput_accumulation_control.gd — disable accumulated input

Hitboxes

NeedAction
Frame-perfect hitMANDATORY exclusively direct_hitbox_query.gd
Volume authoring helperhitbox_component.gd for shapes/layers — never area_entered / get_overlapping_areas for resolution
Proximity guardQuery expanded shape before active frames land

Combos / cancels

NeedAction
Damage scaling ~10%/hitTrack in combo state; store cancel hierarchy on Attack Resources
StatesIDLE/ATTACKING/HITSTUN/BLOCKSTUN… with integer state_frame — peer godot-state-machine-advanced

Netcode

NeedAction
SnapshotsMANDATORY rollback_state_serializer.gd
TransportUDP/ENet via raw_byte_network_sync.gd; peer godot-multiplayer-networking

Balance Guidelines

ElementGuideline
Health10,000-15,000 for ~20 second rounds
Combo damageMax 30-40% of health per touch
Fastest moves3-5 frames startup (jabs)
Slowest moves20-40 frames (supers, overheads)
Throw rangeShort but reliable
Meter gainFull bar in ~2 combos received

For roster / matchup simulation, use godot-monte-carlo-balancer.

Common Pitfalls

PitfallSolution
Infinite combosHitstun decay + gravity scaling
Area2D signal hitsReplace with direct_hitbox_query.gd
Lag input dropsBuffer 8+ frames
DesyncDeterministic loop + rollback serializer

Expert knowledge (on demand)

LLM-ignorance rule: If a general agent would not know it before reading, load the reference — never delete expert deltas.

Reference

Progressive disclosure: open Official Documentation links only when researching a specific API; load Related Skills when routing work to a peer domain — do not preload the whole lattice.

Official Documentation

  • Idle and physics processing — Fighting logic must run on a fixed physics tick (or custom frame counter), not variable _process deltas.
  • Input — Disable use_accumulated_input and sample actions per frame so buffers and motion windows stay deterministic.
  • Input examples — Action maps and event handling patterns behind 5–10 frame buffers and motion detection.
  • Controllers, gamepads, and joysticks — Stick deadzones and digital gate thresholds for clean QCF/DP direction history.
  • Using Area2D — Hitbox/hurtbox volumes, monitoring vs monitorable, and why signal lag is often too late for frame-perfect trades.
  • Physics introduction — Collision layers/masks for High/Low/Throw filtering instead of string groups in the hot path.
  • PhysicsDirectSpaceState2D — Immediate intersect_shape hit resolution without waiting on Area2D overlap signals.
  • PhysicsShapeQueryParameters2D — Shape query setup (mask, exclude, collide_with_areas) for direct hitbox checks.
  • AnimationMixerADVANCE_MANUAL / callback mode so attack clips advance in lockstep with integer frame data.
  • Resources — Store startup/active/recovery, cancels, and balance profiles as .tres data—not hardcoded script constants.
  • High-level multiplayer — Authority, RPCs, and peer roles when adding netplay around a deterministic fighter sim.
  • ENetMultiplayerPeer — UDP/ENet transport for rollback-friendly input exchange without TCP head-of-line blocking.

Related Skills

Prerequisites
  • godot-project-foundations — Physics tick rate, input map, and project defaults must be locked before frame-data systems stay deterministic.
  • godot-input-handling — Action sampling, device mapping, and buffer-friendly input plumbing under motion commands and cancels.
  • godot-2d-physics — Layers/masks and direct space queries are the substrate for hitbox/hurtbox resolution without Area signal lag.
  • godot-characterbody-2d — Grounded movement, facing, and teleport/reset_physics_interpolation contracts fighters still need outside pure hit detection.
Complements
  • godot-combat-system — Shared DamageData / hit confirm patterns that fighting frame data specializes into startup-active-recovery windows.
  • godot-animation-player — Hitbox enable tracks, cancel frames, and recovery locks driven from animation rather than free-running timers.
  • godot-state-machine-advanced — IDLE/ATTACKING/HITSTUN/BLOCKSTUN FSMs with integer state_frame counters instead of await-based recovery.
  • godot-resource-data-patterns — Move lists, cancel tables, and FighterBalanceProfile resources with safe duplication per fighter instance.
  • godot-signal-architecture — Hit confirm, round end, and HUD events without coupling the sim loop to presentation nodes.
  • godot-multiplayer-networking — Peer sync, RPC discipline, and authoritative validation around rollback or delayed-input netcode.
  • godot-adapt-single-to-multiplayer — Prediction, reconciliation, and lobby/late-join patterns when a local fighter becomes netplay-ready.
Downstream / consumers
  • godot-monte-carlo-balancer — Simulate matchup matrices, damage scaling, and punish windows across the roster instead of guessing from AFK→pro PvE bands.
Master
  • godot-master — Library router and mirrored module entry for discovering fighting peers and syncing shared script mirrors.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Godot Genre Fighting AI skill do?

Expert blueprint for fighting games including frame data (startup/active/recovery frames, advantage on hit/block), hitbox/hurtbox systems, input buffering (5-10 frames), motion input detection (QCF, DP), combo systems (damage scaling, cancel hierarchy), character states (idle/attacking/hitstun/blockstun), and rollback netcode. Based on FGC competitive design. Trigger keywords: fighting_game, frame_data, hitbox_hurtbox, input_buffer, motion_inputs, combo_system, rollback_netcode, cancel_system, advantage_frames.

Why use Godot Genre Fighting on TypingMind?

Because you install it once and use it with any model. Godot Genre Fighting is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Godot Genre Fighting in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/thedivergentai/GD-Agentic-Skills/tree/main/skills/godot-genre-fighting. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Godot Genre Fighting?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Godot Genre Fighting?

As many as you like. As long as a model supports skills, you can use Godot Genre Fighting with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Godot Genre Fighting AI skill free?

Yes. It is published on GitHub by thedivergentai under the LGPL-3.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇