Skip to main content
You are viewing: Quick Reference - Common mixin usage patternsSee also: 17 Detailed Mixin Specifications (under Specifications → Tool Mixins)
Import: from gaia.agents.tools.file_monitor_tools import FileToolsMixin

Purpose: Reusable tool sets that can be mixed into your agents. Available Mixins: For complete API specifications of all 17 mixins, see Specifications → Tool Mixins

Composable mixins (registry)

The agent registry exposes these mixins by name in KNOWN_TOOLS (src/gaia/agents/registry.py). A custom agent can compose any of them by listing the tool name — BuilderAgent’s template wires the import and base class for you. Each row is the exact name → module.Class mapping the registry resolves: The sections below show usage patterns for the most common mixins; the file_io mixin is covered under Code Development.

9.1 File Watching Mixin

Detailed Spec: FileToolsMixin The chat-tools FileToolsMixin is narrowly scoped to watching directories for new documents; it does not register read/write/list tools. Use FileSearchToolsMixin (§9.4 below) when you need generic filesystem I/O.

9.2 RAG Tools Mixin

Import: from gaia.agents.tools.rag_tools import RAGToolsMixin Detailed Spec: RAGToolsMixin

9.3 Shell Commands Mixin

Import: from gaia.agents.tools.shell_tools import ShellToolsMixin Detailed Spec: ShellToolsMixin

9.4 File Search Mixin

Import: from gaia.agents.tools.file_tools import FileSearchToolsMixin Detailed Spec: FileSearchToolsMixin FileSearchToolsMixin is the general-purpose filesystem toolbox — it’s what powers read_file / write_file in the chat agent.

9.5 Stable Diffusion Mixin

Import: from gaia.sd import SDToolsMixin Detailed Guide: Image Generation Guide Source: src/gaia/sd/mixin.py
Available Tools:
  • generate_image(prompt, model, size, steps, cfg_scale, seed) - Generate images with SD
  • list_sd_models() - Show available models
  • get_generation_history(limit) - See generated images from session
Supported Models:
  • SDXL-Turbo (default) - Fast, 512x512, ~17s
  • SD-Turbo - Faster, 512x512, ~13s
  • SDXL-Base-1.0 - Photorealistic, 1024x1024, ~9min
  • SD-1.5 - Legacy, 512x512, ~88s

9.6 Vision Language Model Mixin

Import: from gaia.vlm import VLMToolsMixin Source: src/gaia/vlm/mixin.py
Available Tools:
  • analyze_image(image_path, focus) - Get detailed image description
  • create_story_from_image(image_path, story_style) - Generate creative narrative
  • answer_question_about_image(image_path, question) - Answer specific questions
Focus Options: composition, colors, mood, details, all (default) Story Styles: whimsical, dramatic, educational, adventure, any (default)