Locally clears the arcade chat history. Useful for starting fresh or when you want to reset your conversation context.
/clear
Local AI Architecture Deep Dive / Part 4
A collection of slash commands for the Workshop editor. Use these commands to manage your games, edit files, and interact with the AI assistant effectively.
Locally clears the arcade chat history. Useful for starting fresh or when you want to reset your conversation context.
/clearTriggers a Deep Reasoning session for high-complexity architectural tasks. Reuses the edit command logic to target resolution while allowing AI to take over with the new mode.
/deepOpens the editor to the active game, reads its contents, and lets AI decide what changes are needed. Specialized for Workshop file editing with surgical patching support.
/edit
Searches the Workshop editor for a specific string and navigates to it.
Usage: /find <text>. Shows line numbers when found.
/find <text>Specialized protocol for rapid bug fixing. Uses a token-based system to identify target files and surgical edits, allowing quick resolution of issues without full file replacement.
/fix <title> <description>Displays a dynamic list of all registered commands. Shows descriptions and usage examples for each command.
/helpTriggers the AI to brainstorm 3 unique game ideas based on a genre or topic. Useful for generating new concepts when stuck.
/idea <genre>The central orchestrator for all Arcade commands. Handles command parsing, stacking (e.g., /edit /fix), and routing to the appropriate handler based on user input.
/managerRequests a detailed pseudocode planning block from the AI before any implementation. Uses the REASONING_ORCHESTRATOR_V1 protocol to ensure structured thinking and avoid common pitfalls.
/plan <topic>Handles Workshop Arcade publish actions. If called without arguments, it triggers a random game generation; with arguments, it lets the AI handle publishing. Supports markdown code blocks for file content.
/publishRequests a complete replacement for the active Workshop editor file. Uses surgical patching to apply targeted changes while maintaining context from the original file.
/rewrite <title> <description>Flushes VRAM by unloading models from the local inference server. Useful for freeing up memory when you're working with large models or have low RAM. Requires bridge communication to function properly.
/vram