Major Innovation in Agent Skills: Anthropic Upgrades Evals System, Reviving Old Skills (2026-03-09)
Major Innovation in Agent Skills! Anthropic Upgrades Skill Factory with Nuclear-Level Evals System, Developers: Old Skills Revived
Intelligent Gorilla AI Compilation | Editor: Xixi
In the field of AI agents, if you have used Agent Skills, you will definitely be familiar with skill-creator, a no-code building tool for skills released by Anthropic in 2025.
On March 3, Anthropic's official blog released a significant update Improving skill-creator: Test, measure, and refine Agent Skills.
01 - Review of Agent Skills
In October 2025, Anthropic officially launched Agent Skills, a modular and reusable skill package system.
Five Highlights of This Upgrade
- Evals (Automated Evaluation)
- Benchmark Mode
- Multi-Agent Parallel Execution
- Comparator (Blind Comparison)
- Description Tuning (Trigger Description Optimization)
02 - No Reason Not to Install
Anthropic's update to skill-creator has quickly sparked heated discussions among AI agent practitioners and developers.
03 - CI/CD Moment for AI Agents
Anthropic's upgrade to skill-creator essentially brings the most mature testing-benchmark-iteration loop from software engineering to ordinary users and enterprise teams with low barriers.

