feat: Add security assessment module for AWS Bedrock Agent Registry (preview)

[vivekmittal514]

Summary

AWS Bedrock Agent Registry is a new service (currently in preview) that provides a centralized registry for Bedrock Agents. As the service becomes generally available, this framework should include a dedicated security assessment module for it.

Proposed Checks (non-exhaustive)

• Registry access controls and IAM policies
• Agent versioning and aliasing security configuration
• Cross-account agent sharing permissions
• Audit logging enablement for registry operations
• Encryption configuration for registry metadata

Files Likely Touched

• New directory: aiml-security-assessment/functions/security/agent_registry_assessments/ (following the pattern of existing assessors like agentcore_assessments/)
• aiml-security-assessment/template.yaml — add new Lambda function resource and IAM permissions
• aiml-security-assessment/statemachine/ — add parallel execution step for agent registry assessor
• docs/SECURITY_CHECKS.md — add new Agent Registry section

Acceptance Criteria

• New agent_registry_assessments Lambda following the existing module structure in DEVELOPER_GUIDE.md
• At least 5 security checks covering IAM, logging, and access controls
• Integrated into Step Functions parallel execution alongside existing assessors
• Documented in docs/SECURITY_CHECKS.md

Notes

This is a preview-phase service — implementation should track API stability. A draft PR to discuss the check design is welcome before the API is finalized. Follow DEVELOPER_GUIDE.md for the module pattern.