|
Work item:
|
X.sg-AIAof
|
|
Subject/title:
|
Security guidelines for cloud-side AI agent orchestration framework
|
|
Status:
|
Under study
|
|
Approval process:
|
TAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2028-Q2 (Medium priority)
|
|
Liaison:
|
ISO/IEC JTC 1/SC 27, ISO/IEC JTC 1/SC 42, ITU-T SG13, ITU-T SG21
|
|
Supporting members:
|
China Telecom, China Unicom
|
|
Summary:
|
The AI agent orchestration framework has become the core infrastructure for the large-scale implementation of AI agent technology, which realizes agent task decomposition, workflow logic orchestration, multi-tool unified scheduling, multi-agent collaborative management, and execution link closed-loop control. However, the framework faces severe full-stack security risks from infrastructure layer, foundation model layer, agent layer, orchestration and runtime layer to application layer, and the existing international standards have significant gaps in this field.
This recommendation establishes a universal security framework corresponding to the core architecture of the AI agent orchestration framework, and specifies detailed security requirements for each layer of the framework targeting the core security risks of the architecture. This recommendation fills the gap in international standards for AI agent orchestration framework security, and promotes the secure and sustainable development of the global AI agent ecosystem.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2026-06-10 08:36:19
|
|
Last update:
2026-06-10 16:16:07
|
|
