Hosted Agent Explanation Capability Added in skyhook-io/radar v1.13.3

The skyhook-io/radar project has released version v1.13.3, introducing an optional capability for hosted agents to explain saved assessments. According to the project's GitHub release notes, this update allows hosted agents that implement the assessment-bound, tool-free explanation feature to expose the existing "Explain simply" interaction. Hosts can opt in by declaring support via the AgentInfo.assessmentExplanations flag. Unsupported hosted agents will continue to hide explanation generation unless a saved explanation already exists. The change is additive and does not alter local-agent behavior or hosted Apply restrictions.

Operators should verify whether their hosted agents support the explainAssessment request contract before upgrading, as this capability requires backend implementation to function. Unsupported hosted agents will not automatically gain this feature, and the UI behavior depends on the agent's declared capability. Additionally, operators should test the frontend integration in their environment to ensure compatibility with existing workflows and confirm that saved explanations replay correctly. This aligns with a broader pattern in AIOps tools where optional features are gated by capability flags, allowing incremental adoption without disrupting existing functionality.