media MarkTechPost · 1h ago · releases

Microsoft's SkillOpt enables agent skill transfer between Codex and Claude Code

from English

Microsoft, Shanghai Jiao Tong University, Tongji University, and Fudan University researchers developed SkillOpt, a text-space optimizer that trains natural-language skill documents while keeping the target model frozen. The system uses an optimizer model to propose bounded edits based on scored rollouts, exporting the result as a single `best_skill.md` file.

Skills trained on GPT-5.4 transfer to smaller variants with retention rates ranging from 16% to 82% of in-domain gains.
A skill optimized in Codex lifted Claude Code's SpreadsheetBench score from 22.1 to 81.8, slightly exceeding the 80.4 achieved by training directly on Claude Code.
Cross-harness transfer for LiveMath was asymmetric, with only 10% retention when moving from Codex to Claude Code.
The artifact size ranges from 379 to 1,995 tokens, built from one to four accepted edits, and adds zero inference-time calls at deployment.

The shared `best_skill.md` format allows procedural skills to move across different tool APIs, enabling optimization in the cheapest environment for deployment in another.

Importance 2/3 New harness with differentiators MarkTechPost Microsoft Research AI agents Research paper

Related articles

arxiv arXiv cs.LG · 22d ago

REGRIND learns dexterous manipulation from a single human demonstration

The authors present REGRIND, a minimalist retargeting-guided reinforcement learning pipeline that learns dexterous manipulation policies from a single human demonstration. The method retargets human hand-object motion to robot references while preserving spatial and contact relationships, then trains a residual RL policy in simulation to track object-centric keypoints.

arxiv arXiv cs.AI · just now Live

Argus: A General-Purpose Agentic Runtime for Long-Horizon Reasoning

Researchers present Argus, a persistent, self-evolving agentic runtime designed for long-horizon reasoning that separates stable user intent from operational objectives. The system utilizes Manager, Planner, Engineer, and Reviewer roles to execute bounded missions over durable project state, allowing for autonomous execution between operator-owned escalation points.

arxiv arXiv cs.LG · 2d ago · 4 views

Qwen releases Qwen-CUA, a 397B-A17B native computer-use agent

Qwen introduces Qwen-CUA, a native computer-use agent built on a 397B-A17B mixture-of-experts backbone that operates via screenshots and input events without relying on DOM trees or accessibility metadata. The system utilizes a scaffold to maintain up to 20 active screenshots and was trained using a cloud rollout fleet with nearly 100,000 vCPUs across approximately 40,000 verifiable tasks.

arxiv arXiv cs.CL · 2d ago · 2 views

Qwen releases Qwen-CUA, a 397B-A17B native computer-use agent

Researchers have introduced Qwen-CUA, a native computer-use agent built on a 397B-A17B Qwen mixture-of-experts backbone. The system operates by observing screenshots and acting through keyboard and mouse events, without relying on DOM trees or accessibility metadata.

lab Microsoft Research Blog · 2d ago · 8 views

Microsoft Research releases Orchard framework for scalable agentic AI

Microsoft Research has released Orchard, an open-source framework designed to enable scalable and cost-effective research into autonomous agents. The system centers on Orchard Env, a Kubernetes-based environment service that allows researchers to train and evaluate agents across diverse domains like software engineering, web navigation, and personal assistance without rebuilding infrastructure.