China Academy of Information and Communications Technology Launches First Batch of Test Evaluations for Trusted AI Skills Governance, Focusing on Three Core Dimensions

The China Academy of Information and Communications Technology recently launched the first batch of test evaluations for trusted Skills governance. As a key bridge connecting AI agents to the external world, Skills encapsulate various interfaces, tools, and other resources into standardized modules, serving as critical infrastructure for the AI agent ecosystem.

With the rapid development of the Skills ecosystem, credibility issues such as execution deviations, high costs, and scheduling disorders have gradually emerged. This evaluation covers three categories—'execution trustworthiness', 'scheduling controllability', and 'compliance reliability'—with 15 indicators and over 70 sub-items, focusing on functional effectiveness, call management, compliance and security, etc.

The first batch of evaluations covers areas like Skills governance, Skills library governance, and AI agent Skills governance. Material review, technical testing, and expert evaluation are expected to take place in May-June 2026, with final results to be released at a relevant industry conference in July 2026.