- Run main agent:
.venv/bin/python agent.py - Run single example:
.venv/bin/python examples/01_simple_chat.py - Run example via script:
./examples/EJECUTAR_EJEMPLOS.sh 2 - No tests/linting configured in this project
- Imports: Use
from mcp_agent import FastAgent, RequestParams - Async: Always use
async deffor agents withasyncio.run(main()) - Docstrings: Module-level triple-quoted docstrings at top of each file
- Decorators: Use
@fast.agent()with named agents and instruction parameter - Indentation: 4 spaces (not tabs)
- Strings: Use triple-quoted strings for multi-line instructions
- Comments: Inline comments use
#for explanations
- Models: Use prefix
generic.for Ollama models (e.g.,generic.qwen3:4b) - Config file:
fastagent.config.yamlin project root - RequestParams: Use
maxTokensandtemperatureto control output