Implement provider connection tests and health diagnostics #37
Description
Problem
The operator needs a reliable way to test whether each provider can actually start, stream, and resume before real work is attempted.
Scope
- Define connection-test flows for all supported providers
- Cover launch, stream, resume, tool access, and auth diagnostics at a planning level
- Ensure failures are attributable and explainable
Out of scope
- Long-running session telemetry dashboards
- Generic settings layout concerns unrelated to diagnostics
Implementation notes
- Make health diagnostics consistent across providers
- Ensure the issue covers both success and failure reporting
- Reuse the Toolchain Center as the operator entry point
Definition of Done
- The issue defines what a successful connection test proves for each provider
- Health and failure states are concrete enough to implement later
Verification
- Review the issue against the provider adapter and telemetry epics