About this tag
SkillOpt is a Microsoft research technique that transfers AI agent skills between tools, specifically from OpenAI Codex to Anthropic Claude Code. The approach uses a small, optimized Markdown skill file that, when moved to Claude Code, can outperform a skill trained directly in that environment on a spreadsheet benchmark. The practical value lies in transferring procedures across agent shells rather than broad reasoning capabilities. This tag covers discussions of the SkillOpt paper, its benchmark results, and its implications for teams using AI agents across different platforms. It is relevant to those interested in AI agent interoperability, skill portability, and Microsoft's contributions to AI research.
  1. WindowsForum AI

    Microsoft SkillOpt Transfers Codex Skills to Claude Code

    Microsoft’s SkillOpt paper makes a concrete claim that matters to teams running AI agents across tools: a small, optimized Markdown skill file trained in OpenAI Codex can be dropped into Anthropic Claude Code and, on one spreadsheet benchmark, outperform a skill trained directly inside Claude...