https://blog.chameleoncloud.org/authors/agustin-leon/Recent content in Agustin Leon on Blog | ChameleonHugoen-usSun, 26 Apr 2026 00:00:00 +0000https://blog.chameleoncloud.org/posts/openmcp-reproducible-benchmarking-mcp-agents/Sun, 26 Apr 2026 00:00:00 +0000https://blog.chameleoncloud.org/posts/openmcp-reproducible-benchmarking-mcp-agents/<p><em>Agustin Leon and Anup Raj Niroula, New York University</em></p> <hr> <h2>What is the central challenge your experiment investigates?</h2> <p>Computer-use agents are language model programs that operate a computer through the same channels a person does: controlling a cursor, typing into applications, and calling tools. Think of it like an assistant that can organize your Notion notes, update a GitHub issue, or manage your calendar without needing step-by-step instruction. Through <a href="https://www.anthropic.com/news/model-context-protocol">Model Context Protocol (MCP)</a>, an open standard released in late 2024, agents can perform actions with specific applications in a consistent way. Adoption of MCP has been rapid: by the end of 2025, PulseMCP, a popular directory of MCP servers, had already <a href="https://www.pulsemcp.com/statistics">tracked over 15 million downloads</a> and more than 6,000 MCP servers.</p>