[Feat]: Enhancing A2A Authorization for Agent Registries with Fine-Grained Control

### Is your feature request related to a problem? Please describe.

The current Agent-to-Agent (A2A) authorization guidelines are too high-level, failing to adequately address the complex authorization needs of modern enterprise environments. As more organizations migrate to agent registry models for A2A discovery, they face challenges in implementing granular control over agent interactions. Existing authorization mechanisms, limited to skills, actions, and OAuth scopes, are insufficient for scenarios requiring more precise policy configurations.

### Describe the solution you'd like

We are requesting an expansion of the A2A authorization implementation guidelines to include a **fine-grained authorization framework**. This framework should provide concrete scenarios and guidance for configuring policies based on **agent containers**. By enabling the logical grouping of agents using attributes within these containers, and then defining policies beyond the current scope limitations, enterprises can achieve a higher level of control and security. This would allow agents to operate within clearly defined boundaries, even as authorization implementations vary based on specific business requirements.

### Describe alternatives you've considered

_No response_

### Additional context

_No response_

### Code of Conduct

- [x] I agree to follow this project's Code of Conduct