godot-interactive
Interact with a running Godot game via MCP — launch, screenshot, click, inspect scene tree, get/set properties. Powered by godot-mcp + GodotMCPBridge autoload.
What this skill does
# Godot Interactive Testing Skill
Playwright-like interaction loop for Godot games via MCP. Launch, observe, click, inspect, and manipulate a running game through a bridge autoload.
## Overview
This skill enables LLM-driven interactive testing of Godot games through the godot-mcp MCP server and a GodotMCPBridge autoload. You can:
- Launch and stop Godot projects via MCP
- Capture viewport screenshots automatically or on-demand
- Inspect the scene tree with Control node positions and text
- Simulate clicks, key presses, and input actions
- Read and write node properties at runtime
- Implement iterative test/debug loops (observe → inspect → interact → observe)
This is particularly useful for:
- Automated UI testing
- Debugging visual issues
- Verifying game state without manual playtesting
- Creating reproducible test scenarios
- Validating UI layouts and interactions
## Prerequisites
**godot-mcp server:**
- Installed at `~/code/godot-mcp` (or configurable path)
- Built with `npm install && npm run build`
- Repository: https://github.com/yourusername/godot-mcp
**Claude Desktop configuration:**
- MCP server configured in `~/.claude/mcp.json` (see Setup)
**Godot project setup:**
- GodotMCPBridge autoload installed
- Remote debugger enabled (automatic when running via `run_project`)
## Setup for New Projects
### 1. Copy the Bridge Autoload
Copy the bridge script into your project:
```bash
cp assets/templates/godot_mcp_bridge.gd res://scripts/autoload/godot_mcp_bridge.gd
```
### 2. Register the Autoload
Add to `project.godot`:
```ini
[autoload]
GodotMCPBridge="*res://scripts/autoload/godot_mcp_bridge.gd"
```
Or via Godot editor:
- Project → Project Settings → Autoload
- Path: `res://scripts/autoload/godot_mcp_bridge.gd`
- Name: `GodotMCPBridge`
- Enable: Yes
### 3. Create MCP Configuration
Create `.mcp.json` at project root (use template):
```bash
cp assets/templates/mcp.json.template .mcp.json
```
Edit paths:
```json
{
"mcpServers": {
"godot": {
"command": "node",
"args": ["<PATH_TO_GODOT_MCP>/build/index.js"],
"env": {
"GODOT_PATH": "<PATH_TO_GODOT_BINARY>"
}
}
}
}
```
**Example paths:**
- macOS: `/Applications/Godot 4.5.app/Contents/MacOS/Godot`
- Linux: `/usr/bin/godot` or `~/godot/bin/godot`
- Windows: `C:/Godot/godot.exe`
**Project-local godot-mcp:**
If you keep godot-mcp in your project (e.g., `.tools/godot-mcp`):
```json
"args": [".tools/godot-mcp/build/index.js"]
```
## MCP Tools Reference
All tools are provided by the godot-mcp server and work through the GodotMCPBridge autoload.
### run_project(projectPath)
Launch a Godot project with remote debugger enabled.
**Parameters:**
- `projectPath` (string): Absolute path to project directory containing `project.godot`
**Returns:**
- Success confirmation with process ID
**Example:**
```typescript
run_project({ projectPath: "/Users/ben/code/my-game" })
```
**Notes:**
- Automatically enables remote debugger on port 6007
- Takes 2-3 seconds for project to fully start
- Bridge begins auto-capturing screenshots every 2 seconds
- Check `get_debug_output()` if game doesn't appear to start
### stop_project()
Stop the currently running Godot project.
**Parameters:** None
**Returns:**
- Success confirmation
**Example:**
```typescript
stop_project()
```
**Notes:**
- Kills the Godot process gracefully
- Cleans up debugger connection
- Safe to call even if no project is running
### get_debug_output()
Retrieve stdout/stderr from the running Godot process.
**Parameters:** None
**Returns:**
- Array of log lines (most recent first)
**Example:**
```typescript
get_debug_output()
// Returns: ["Frame 1234", "Button pressed", "MCP command: click", ...]
```
**Use cases:**
- Check if game started successfully
- Debug print statements from your game
- Verify bridge initialization ("Godot MCP Bridge initialized")
- Inspect error messages
### game_screenshot()
Capture the current viewport and return as base64 image.
**Parameters:** None
**Returns:**
- Base64-encoded PNG image of the game viewport
**Example:**
```typescript
game_screenshot()
// Returns: base64 PNG data
```
**Notes:**
- Captures the main viewport at full resolution
- Bridge auto-saves screenshots to `/tmp/godot_screenshot.png` every 2 seconds
- This tool provides immediate capture via MCP response
- Works on stock Godot (unlike `capture_screenshot` from upstream DAP)
### game_scene_tree(path, depth)
Inspect the scene tree with node types, positions, and properties.
**Parameters:**
- `path` (string, optional): Root path to start inspection, default: `/root`
- `depth` (number, optional): Traversal depth, default: 3
**Returns:**
- JSON tree structure with nodes
**Node properties:**
- `name`: Node name
- `type`: Node class (Control, Button, Label, Node3D, etc.)
- `rect`: For Control nodes: `{x, y, w, h}` in screen coordinates
- `visible`: For Control nodes: visibility state
- `text`: For Button/Label nodes: display text
- `children`: Array of child nodes (up to depth limit)
**Example:**
```typescript
game_scene_tree({ path: "/root/Main/UILayer", depth: 2 })
// Returns:
{
"name": "UILayer",
"type": "CanvasLayer",
"children": [
{
"name": "VictoryPopup",
"type": "CenterContainer",
"rect": {"x": 640, "y": 360, "w": 400, "h": 200},
"visible": true,
"children": [
{
"name": "OKButton",
"type": "Button",
"rect": {"x": 720, "y": 480, "w": 80, "h": 40},
"text": "OK",
"visible": true
}
]
}
]
}
```
**Use cases:**
- Find clickable UI elements (buttons, controls)
- Verify UI layout and positioning
- Check visibility states
- Locate nodes by name/type for property access
**Key gotcha:** UI coordinates are full resolution (e.g., 1920x1080), not screenshot pixel coordinates. Always use `game_scene_tree` to get accurate `rect` values for clicking.
### game_click(x, y, button)
Simulate a mouse click at screen coordinates.
**Parameters:**
- `x` (number): Screen X coordinate
- `y` (number): Screen Y coordinate
- `button` (number, optional): Mouse button index, default: 1 (left click)
- 1: Left button (MOUSE_BUTTON_LEFT)
- 2: Right button (MOUSE_BUTTON_RIGHT)
- 3: Middle button (MOUSE_BUTTON_MIDDLE)
**Returns:**
- Success confirmation with coordinates
- Automatically captures screenshot after click
**Example:**
```typescript
game_click({ x: 720, y: 480, button: 1 })
// Returns: {"type": "click", "success": true, "x": 720, "y": 480}
```
**Notes:**
- Simulates press + release with 50ms delay
- Takes a screenshot 200ms after release
- Use `game_scene_tree` to find button `rect` coordinates
- Coordinates are in full window resolution, not SubViewport pixels
**Workflow:**
1. Call `game_screenshot()` to see the game
2. Call `game_scene_tree()` to find button positions
3. Call `game_click()` with button's center coordinates
4. Call `game_screenshot()` to verify result
### game_key(key)
Simulate a key press using Godot key names.
**Parameters:**
- `key` (string): Godot key name (e.g., "Space", "Escape", "Enter", "A")
**Returns:**
- Success confirmation
- Automatically captures screenshot after key press
**Example:**
```typescript
game_key({ key: "Space" })
game_key({ key: "Escape" })
game_key({ key: "Q" })
```
**Common key names:**
- Letters: "A", "B", "C", etc.
- Numbers: "0", "1", "2", etc.
- Special: "Space", "Enter", "Escape", "Tab"
- Arrows: "Up", "Down", "Left", "Right"
- Function: "F1", "F2", etc.
**Notes:**
- Uses `OS.find_keycode_from_string()` - must be valid Godot key name
- Simulates press + release with 50ms delay
- Takes screenshot 200ms after release
- Case-sensitive for letters (use uppercase "A", not "a")
### game_action(name)
Trigger a Godot input action defined in InputMap.
**Parameters:**
- `name` (string): Action name from project InputMap
**Returns:**
- Success confirmation
- Automatically captures screenshot after action
**Example:**
`Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.