Tool Implementation Patterns

Victor supports two tool styles:

• Class‑based tools (BaseTool)
• Function‑based tools using the @tool decorator

Recommendation

• Use @tool for new tools.
• Migrate class‑based tools when you’re already touching them.

Minimal @tool Example

from victor.tools.decorators import tool

@tool
async def my_tool(param: str) -> dict:
    """Short description of the tool."""
    return {"success": True, "result": param}

Notes

• Both styles are supported during migration.
• Keep tool docstrings concise; they become tool descriptions.
• Regenerate the catalog after changes: python scripts/generate_tool_catalog.py