Exonfang/gdllm-godot-agentic-harness
A plugin that lets AI agents code, edit scenes, and debug directly inside the Godot game engine
GDLLM embeds a chat panel into the Godot game engine editor, letting large language model (LLM) agents read and write files, edit scenes, and run and debug games as an open-source plugin. Every action the agent takes is fully visible and saved in session history. The project claims it completes the same tasks using roughly half the tokens compared to Opencode, another agentic tool.
What it does
- Connects to any LLM provider's API, including OpenAI, Anthropic, and Ollama, and can also run on a ChatGPT Plus/Pro subscription without separate API billing
- Uses 'progressive disclosure' to pull in only the information needed at each moment, keeping token usage low, while offering a wide toolset for reading/writing files, editing scenes, running and debugging games, and delegating to subagents
- Automatically validates code edits through the engine and surfaces errors back to the agent, plus includes loop-detection that stops an agent when it's thrashing on the same action
- Supports per-session permission gates (read-only, make changes, make changes and delete) and can restrict tool calls to stay within user and project directories for safety
- Written in GDScript, with an optional MarkdownLabel addon that prettifies the model's text output
Why it matters
For game developers who want to hand off repetitive coding, scene setup, or debugging to AI while still seeing exactly what it's doing, this offers a transparent, in-editor option. Its token-saving design also translates directly into lower API costs for solo developers or small teams.
Terms in this repo
- agentic harness · a framework that wraps an LLM so it can autonomously call tools to complete tasks
- token · the basic unit LLMs process text in, and the basis for API billing
- progressive disclosure · a design that reveals only the information needed at each step, keeping initial context small
- MCP server · a standardized server that gives AI models access to external tools and data
- subagent · a secondary agent that the main agent delegates specific tasks to
Repository description (English)
A fully transparent in-editor agentic harness for large language models inside the Godot Engine. Connect to any LLM provider via API and provide your agents the tools, context management, and guidance to effectively complete most tasks in Godot. Every action the agent takes is fully surfaced and persists in session history.
Open on GitHubTrending repos
- vorssaint/vorssaint-utilsOne free menu bar app replaces a dozen paid Mac utilities
- Alishahryar1/free-claude-codeA local proxy that lets coding AI agents run on 49 free or cheap model providers instead of one paid service
- freestylefly/awesome-gpt-image-2A library of 532 reverse-engineered prompts that turn GPT-Image2 into a predictable image-making tool
- block/buzzAn open-source workspace where humans and AI agents chat, code, and review in the same rooms
- NousResearch/hermes-agentNous Research's Hermes is an AI agent that gets smarter the more you use it
- virgiliojr94/book-to-skillA tool that turns technical book PDFs into on-demand reference skills for AI coding agents
- VoltAgent/awesome-agent-skillsA single hub collecting over 1000 'how-to' manuals that make AI coding assistants act like experts
- anthropics/claude-plugins-communityA shared shelf where anyone's Claude add-ons get listed for install