Francisco Liberal

AI Integrations Engineer

Francisco is a full-stack developer and .NET expert who builds the integrations that make AI agents actually useful. At Arcade, he focuses on creating robust connectors that bridge the gap between AI capabilities and real-world applications. His obsession with quality and efficiency ensures that the tools developers rely on just work.

Articles (2)

All authors

Building High-Quality MCP Tools with Arcade.dev Evals

FEBRUARY 26, 2026

MCP

Building High-Quality MCP Tools with Arcade.dev Evals

Tool definitions are not function signatures. They are menu items an LLM reads to decide what to call and how to fill the arguments. This post covers what separates good tool definitions from bad ones, with before/after examples and a practical checklist.

Everything Is a Test: How to Evaluate MCP Tools for Reliable AI Agents

FEBRUARY 26, 2026

MCP

Everything Is a Test: How to Evaluate MCP Tools for Reliable AI Agents

Tool quality directly affects agent reliability. Arcade.dev Evals lets you test MCP tool definitions against LLM behavior in simulated conversations, scoring whether the model selects the right tool and populates arguments correctly, before you deploy to production.

Get early access to Arcade, and start building now.

Docs